Current draw of typical FSAE electrical components [Archive]

jlangholzj

07-22-2013, 09:29 AM

Originally posted by murpia:
<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">Originally posted by Claude Rouelle:
Why should the participants of this forum help you if you do not introduce your self?

Originally posted by Zac:
because he's been an active poster on the forum for years?

Posts: 545 | Registered: July 15, 2003

Sorry Claude, the dangers of posting too briefly in a hurry...

I was thinking about the likely specifications for an FSAE-specific power management module, to replace the usual collection of fuses and relays. Now, this would need to be fairly generic, and suit a range of cars and components. Hence it's not so easy simply to measure the loads for just one car.

To be specific, I was wondering if the majority of the typical loads can be powered through a single 20 size DTM contact, which is rated at 7.5A continuous.

I hope that helps, Ian </div></BLOCKQUOTE>

well..what are your current loads? What's the DTM connector sized at? why would you want to run...say 15 amps through the connector...how COULD you run 15 amps through the connector? Why do you want to make it fit a range of components. How would you make it fit a range of components.

...these all seem like pretty easy questions to me...if you're looking for "verification" of a design, we're going to need a little bit more than just...how do i fit it to multiple loads with this connector.

And yes, it should be fairly easy to measure loads from one car...and make ASSUMPTIONS about some other items. So if you measure all your components and your fan is the highest load drawing, why not make all your circuits capable to handle that much load....what would be the advantages...disadvantages???

Well then....what if you want to use a BIGGER fan? Can you components handle that? What about the PCB....

again...this all seems like a pretty logical step by step process that shouldn't be terribly difficult to follow through with.

MCoach

07-22-2013, 09:35 AM

Ian,

pretty much our entire car is DTM connectors. 7.5A is also just a recommendation. http://fsae.com/groupee_common/emoticons/icon_biggrin.gif
If you check the connector and stator output of the engine, that'll give you an idea of what should be used to for main wire diameter. Other than that, voltage drop calculations are decent to check, but for anything other than the starter motor, you'll want to go larger than the 'optimal' wire size because working with 30 gauge or smaller is a pain and fatigues readily to the high engine vibrations.

Short answer, yeah, I guess. I've pulled all of the car current through a DTM connector (sans starter motor), but I wouldn't recommend it. I wouldn't do it for a long term solution. Maybe a DT or a DTP connector. DTPs are rated at 25 amps.

Good luck.

Fantomas

07-22-2013, 01:22 PM

Given the fact that Ian seems to be already an engineer, I guess, that he plans to build a commercially available power distribution box and ist just looking for data.
If not, he has spend a very looong time in FSAE, at least 10 years...

jlangholzj

07-22-2013, 01:33 PM

Originally posted by Fantomas:
Given the fact that Ian seems to be already an engineer, I guess, that he plans to build a commercially available power distribution box and ist just looking for data.
If not, he has spend a very looong time in FSAE, at least 10 years...

I see the merit in this statement but why buy one for ~$1200 when you can make one for ~$100

I'm guessing that he's getting data to make one...either himself or someone else.

Jay Lawrence

07-22-2013, 09:49 PM

Not sure about the minor components but the starter motors on our CBR600 engines draw about 100A

murpia

07-23-2013, 03:23 AM

Originally posted by Fantomas:
Given the fact that Ian seems to be already an engineer, I guess, that he plans to build a commercially available power distribution box and is just looking for data.
'plans to build' is a bit premature but yes I was interested to understand the current range FSAE users would require from such a device.

Originally posted by jlangholzj:
I see the merit in this statement but why buy one for ~$1200 when you can make one for ~$100
I agree, but why make one for $100 if you could buy one for $200...

Regards, Ian

Kevin Hayward

07-23-2013, 05:25 AM

Ian,

Our sparkies put something together that is very similar to what you are suggesting. It has taken years of fiddling to get it to a pretty simple and relatively problem free state. On the face of it I would be all for a $200 simple PDM. It is not very cost effective to do these things in ultra small runs when you don't have the inhouse capability to manufacture the base circuit boards. We get them manufactured externally and usually end up with heaps of extras because the unit cost and batch cost are not too different.

The PDM from places like Motec has too much unused capability and is to expensive for it to be good value for money. A couple of things that would be worthwhile (and maybe justify a slightly higher price) a couple of simple outputs (or CAN communication) to monitor current draw and board temp.

Feel free to PM me if you want a contact for one of our students who has been working on our board for a while. It is nothing fantastic as the design brief was to be as barebones as possible.

Kev

murpia

07-23-2013, 08:29 AM

Originally posted by Kevin Hayward:
The PDM from places like Motec has too much unused capability and is to expensive for it to be good value for money. A couple of things that would be worthwhile (and maybe justify a slightly higher price) a couple of simple outputs (or CAN communication) to monitor current draw and board temp.
A good discussion is how low-end could a PDM be and still be useful to an FSAE team?

You have to have some form of comms to configure the unit, so it may as well support CAN. Then you need some inputs to turn on and off loads. The advantages over relays & fuses should be additional reliability, the ability to measure your load currents and the ability to have smart-resetting in the event of an intermittent fault that would otherwise simply blow a dumb fuse.

Regards, Ian

jlangholzj

07-23-2013, 09:02 AM

Originally posted by murpia:
<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">Originally posted by Kevin Hayward:
The PDM from places like Motec has too much unused capability and is to expensive for it to be good value for money. A couple of things that would be worthwhile (and maybe justify a slightly higher price) a couple of simple outputs (or CAN communication) to monitor current draw and board temp.
A good discussion is how low-end could a PDM be and still be useful to an FSAE team?

You have to have some form of comms to configure the unit, so it may as well support CAN. Then you need some inputs to turn on and off loads. The advantages over relays & fuses should be additional reliability, the ability to measure your load currents and the ability to have smart-resetting in the event of an intermittent fault that would otherwise simply blow a dumb fuse.

Regards, Ian </div></BLOCKQUOTE>

Why use additional inputs to turn on an off loads when you've already got CAN communications?? If you want a "configurable unit" with a bunch of other fun stuff, you've already got a micro and everything else needed to turn inputs on/off.

I'm slowly heading this way, ours this year was constructed with several N-FET's that had just simple analog inputs to turn the circuits on and off. Hopefully sometime soon I'll have current sensing, PWM and other capabilities on the board.

Kevin Hayward

07-23-2013, 09:22 AM

Ian,

I may alienate some of the more electrically minded on this forum, but I think with the combustion cars at least when it comes to electrics the less the better. I think a very low end device would suit 80-90% of the teams. The big advantage of the device would be putting it all in one small easy to wire to package, not in the communications to and from. The only interesting thing on our simplified board is that we run custom ignition modules on board. Could definitely separate that, but it would be great to have on board. It saves space and money for an item that is way overpriced if you buy the original bike stuff.

Packaging is generally what we are after and it is nice to have one box for the switching and one for the ECU, some sort of driver display and nothing else.

I don't think you would need to run CAN to turn things on. I think a very simple connection to activate the switching would be fine. Generally you are not short of outputs on an FSAE vehicle. The big problem is inputs to the ECU. Anything you put in will take away from the pins available for logging important things. Running an external logger considerably increases both the reported cost of the car and the complexity of the wiring. Being able to use your ECU at comp to take care of the logging duties and keeping to the essentials is a nice fix.

If I was buying I would something a little less configurable and more direct instructions as to what limits etc are okay. Nothing worse than having to learn yet another interface for another device that you just want to plug in and have it work. I would probably want to buy two in the first instance to keep a spare handy. If it was simple enough I think you would have a lot of the more mech. minded people building hillclimb and other motorcycle powered cars pretty interested. I work with a few people locally and they generally just want the electrics to work. They want to buy shifting systems off a vendor and have them install and work immediately. They want to pay someone to do the wiring and have it work first time and every time. Most FSAE teams are in the same boat whether they realise it or not. Electrical failures are one of the biggest problems for these cars finishing events, at least to my observations they outnumber mechanical failures easily.

The main points of interest would be:

- Easy connections to wiring harness
- Small physical size and weight
- Easy way to mount to vehicle with fastening methods such as holes designed in (Motec ECU is annoying here)
- High reliability
- Good installation instructions
- Detailed specs
- Well labelled unit with colour coding as necessary
- Good environment protection

Lesser points:

- Ability to log current draw and temps
- Non physical fuses
- Possibly the ability to purchase sub looms for the fan / pumps / starter motor
- Maybe the ability to purchase a sub loom for connecting to a motec (or other popular ECUs)
- Good suitable recommendations for fan / water pumps

Not that important (to me anyway):

- Ability to program

Also worth noting that in our team at least if this item was at a reasonable cost there would be one for each car. That means unlike an ECU which might be in any of the running cars at a given time. It would have no bonus to us of being able to be flexible enough to suit both a single, a twin, and a four. You may want to consider a few different versions especially if you wanted to include the ignition modules on board.

One last thing in a bit of a brain dump, there is one other item that is pretty annoying to package at that is the voltage regulator. If you could build something like that in to the box it would be even more interesting. The case weight could then be doing something and the temp reading would be telling you how well your regulator is being cooled.

You may even want to think about attaching a master switch. It would massively reduce heavy wire length (and weight) and be a one stop box for most of the electrics. This will lead you to determining where the unit is mounted as it will be dictated by the required master switch location. It depends very heavily on how small you could make the unit. It is is only a couple of times bigger than the starter switch it is no big deal. If it is the size of a lunchbox it would be pretty bad.

Kev

Rex

07-23-2013, 04:46 PM

Originally posted by Claude Rouelle:
Why should the participants of this forum help you if you do not introduce your self?

Originally posted by Zac:
because he's been an active poster on the forum for years?

Wow. I realize this is probably an example of a simple oversight (Claude not seeing Ian's post count/history), but maybe it's also an indication that this whole "judges = forum police" thing has gone too far. If I do a quick search it's easy to see a whole bunch of recent examples of people getting short/condescending/snarky/rude responses from competition officials. Why is that?

There's absolutely no question that officials/judges/etc being on these forums adds value and should be appreciated in general. But I'm not sure how snarky responses or demanding that people maintain proper "internet decorum" by introducing themselves before anyone is allowed to help them is value-added or is a good way to encourage participation and open discussion. This is especially apparent when it's aimed at someone like Ian, who has been on this forum basically since it was founded, has contributed much here, and is probably doing a bit of semi-vague product research that might even benefit the FSAE community overall if he decides to build it.

So in summary, I'm not sure why the recent trend towards snarky judge responses has occurred, and I'm certainly not saying that some of the snarky responses aren't warranted (many of the silly questions on here annoy me too!). However I also think that while being a competition official (or an official of any kind in anything in life) commands huge respect, that type of respect is also a 2-way street and probably includes giving everyone who asks a random question on these forums the benefit of the doubt. If we discourage one curious but ignorant/inexperienced student with a snarky response, that's far worse of a tragedy than allowing 10 people to be lazy or uninformed. We all asked silly questions and/or received more info than we gave when we started our educational journey. Anyway, just food for thought from an old guy who isn't involved in competitions anymore (and therefore realizes my opinion isn't worth much!) but still reads these forums for whatever reason...

Kevin Hayward

07-23-2013, 05:07 PM

Rex,

It is a bit annoying to see it on just about every post. I believe it is better not to post your affiliations or experience. Others should feel free to determine the validity of your comments based on content, not because the words come from an expert.

Kev

Pete Marsh

07-23-2013, 08:08 PM

Ian,
I addition to FSAE I think there is a market with club level motor sport for a really simple PDM. Mechanical relays and push to reset thermal overloads (the old aircraft style) are the favoured set up for many, and maybe you can simulate that behaviour, and convince them your box of tricks is smaller lighter and better.
And good, but cheap electrical connections. The mating plugs can't cost as much or more than the unit! Many people at this level would never consider a bespoke loom with Mil Spec connectors throughout.
I think you should also consider a "factory programmed" unit that is simply plug and play. There are plenty of people that would prefer that.

Do that for $200.00 and I'll buy one.

Pete

murpia

07-24-2013, 01:12 AM

Originally posted by Kevin Hayward:
Rex,

It is a bit annoying to see it on just about every post. I believe it is better not to post your affiliations or experience. Others should feel free to determine the validity of your comments based on content, not because the words come from an expert.

Kev
Hi Kevin,

I have a personal dislike of 'signatures', hence I usually don't bother with one on any forums I participate in (not just FSAE). As you, I find that the content of a comment or post is usually enough to determine validity.

Regards, Ian

jd74914

07-24-2013, 05:13 AM

murpia

07-24-2013, 05:55 AM

Originally posted by Pete Marsh:
Ian,
I addition to FSAE I think there is a market with club level motor sport for a really simple PDM. Mechanical relays and push to reset thermal overloads (the old aircraft style) are the favoured set up for many, and maybe you can simulate that behaviour, and convince them your box of tricks is smaller lighter and better.
And good, but cheap electrical connections. The mating plugs can't cost as much or more than the unit! Many people at this level would never consider a bespoke loom with Mil Spec connectors throughout.
I think you should also consider a "factory programmed" unit that is simply plug and play. There are plenty of people that would prefer that.

Do that for $200.00 and I'll buy one.

Pete
Thank you too to Jim & Kevin for their comments on a 'fit and forget' style PDM.

I think a single product that comes 'factory' configured to simulate a set of traditional relays and fuses, but can be re-configured by advanced users, could work. The purchased config can easily be factory programmed to a default accessory load set, or simply be left at the largest 'virtual' fuse rating that still protects the internal drivers and the connector.

The only internal cost difference between a CAN and non-CAN device is a single transceiver chip, so CAN support is a no brainer. The user can opt not to buy a USB->CAN adapter initially (this would be pretty cheap too, maybe $50) and can buy one later to upgrade if they want to integrate the PDM further.

How many outputs would make sense? Would they all need to be >>10A or would a mix of outputs be OK?

An appropriate power connector is a Deutsch DTF13-12PA (this has 12 contacts). The mating half is $5 plus wire and they are light, sealed, robust, serviceable without tools and reasonably compact. The PDM itself could be made small enough to follow the profile of the connector and be about 60mm long (I estimate).

A single 12pin connector offers 6 outputs turned on by 6 pull-to-ground inputs. These inputs would be compatible with the typical relay driver included in most ECUs like the Megasquirt or the PE. Power should probably be an M8 'stud' like the MoTeC PDMs and a smaller 3-pin DTF13-3P would make a good CAN + signal Gnd connector.

I think a bespoke case would be needed, especially given the M8 stud connection. That might end up being the highest cost item, but essential for reliable operation & good heatsinking in my opinion. I still think $200 is achievable.

Regards, Ian

SNasello

07-24-2013, 07:26 AM

Ian,

I'll throw some numbers at you. Depending on the size of the fan, you are probably looking at anywhere from 3 to 12A for the sizes that would be used in FSAE.

The Davies Craig water pumps which are often used in FSAE can draw up to 5 or 6A.

Fuel Pumps are where the current consumption can vary quite a bit, but normally lie in the range of 6-10A, depending on Fuel Pressure.

Starter Motors are where it starts to get interesting, with peak turn on currents up to 200A and cranking currents between 70A and 100A.

I think its a great idea for an off the shelf, low cost PDM. If the outputs can be PWM controlled you will definitely beat the Motec PDM on one feature.

murpia

07-24-2013, 08:59 AM

Originally posted by SNasello:
Ian,

I'll throw some numbers at you. Depending on the size of the fan, you are probably looking at anywhere from 3 to 12A for the sizes that would be used in FSAE.

The Davies Craig water pumps which are often used in FSAE can draw up to 5 or 6A.

Fuel Pumps are where the current consumption can vary quite a bit, but normally lie in the range of 6-10A, depending on Fuel Pressure.

Starter Motors are where it starts to get interesting, with peak turn on currents up to 200A and cranking currents between 70A and 100A.

I think its a great idea for an off the shelf, low cost PDM. If the outputs can be PWM controlled you will definitely beat the Motec PDM on one feature.
Thanks for the numbers, Stefan. I can definitely say that I do not plan to design a unit capable of controlling a starter motor, though...

PWM control is an interesting topic. To do that requires a half-bridge electrical architecture not a simple high-side drive. You then have to deal with ground recirculation currents, so need a high current ground pin for the load as well as a high current supply pin. I'll have a think about that.

As a general query, how would you as users expect the CAN control to work? I have some experience with the GEMS PM1 which has a very simple CAN interface. It simply takes 'virtual fuse' currents and on / off signals sent by a smart central controller on a dedicated CAN id. It does not include arbitrary CAN decoding of sensor values or have any internal control strategy. Would this be OK for you guys, or were you thinking the central controller functions would be included as well?

Thanks, Ian

jlangholzj

07-24-2013, 09:00 AM

Ian since it seems like you're going to get down and dirty with this..or at least contemplating it (like i do at times :P ) I'll feed you some thoughts.

personally...i absolutely HATE screw terminals. Just as a heads up http://fsae.com/groupee_common/emoticons/icon_razz.gif I do however realize that there's a few spots where they can be useful.

-allow for serial communications to the micro via command terminal like terra term or something similar. This will allow for diagnostics, modification commands and so forth.

-wouldn't be that hard to allow a CAN communicated and an analog input to drive the outputs in the same package. Have a jumper that's 2-way configurable. Position 1 allows inputs from the a-d pins on the micro (which are tied to inputs) and position 2 allows for a CAN-read in that controls the outputs, or other serial communication for that matter...CAN...rs232...etc...ours currently follows the analog input option but i would like to have CAN capability

-current sensing. This is going to be relatively easy...why? Because you should already be using power fet's to drive these...instead of relays. The added bonus to this...is that there's fet's with a built in current sense!! Feed that guy back into the micro and it can serve as a fuse/overprotect.
<pre class="ip-ubbcode-code-pre">
for(I>10){
if(T_on > 5sec){
turn off}
}
</pre>

something similar to that..where I is the sensed current and T_on is how long its been there.

-PWM. Could do this a couple different ways. Obviously using the FET's you've already got a very viable PWM option when coupled with the micro. Either you can program it via the micro (hard set value) or do something like a small rotary pot to adjust values. The former of course allows for much more flexibility. Any output would be able to be configured as a PWM output as all the driving factor would be is the PWM output routine.

another thing to consider is the use of a budget FPGA like a cyloneII instead of just a micro. If you do use just a micro, be careful at how fast you're clocking it...PWM takes a LOT of clock cycles...not physically...but when you look at the wasted cycles...plus the interrupting....its quite the hog on the system.

-flyback. Most fet's have an internal diode for their protection but using that diode as a flyback diode is bad practice. Make sure that you've got flyback protection diodes in the circuit so that your transient spikes dont make something else in your circuit to go poof.

-grounding planes. I'm assuming that you're doing this on a PCB so when you do the board...make sure you've got proper EMI isolation. If your ground planes aren't happy....it can lead to some undesired results with signals like PWM. In our application its probably not that big of a deal...but again, generally good practice.

personally i see this box as an "everything but the starter" box. The draws of 100-150A (PLUS imagine the transients!) of the starter aren't something you're going to want to run through a PCB. even 4oz copper would require ridiculous traces and layers.

Use the 12-pint deutsch connectors and build an SLA box that this will fit into. design retention holes to screw into the bottoms of the connector and then seal the opening with silicone or some potting....just the opening...please don't fill the whole thing...its not fun....don't ask me how i know http://fsae.com/groupee_common/emoticons/icon_razz.gif

This is a seriously good project and these are all things I've been working at for the last year or so. I'm still going to come back to my original comment and say that all of these ratings you're going to be able to easily find. Almost all electronics manufactures will have steady-state current loads...etc...Look around at a bunch of options and it should give you a good idea http://fsae.com/groupee_common/emoticons/icon_wink.gif

jlangholzj

07-24-2013, 09:05 AM

Originally posted by murpia:
PWM control is an interesting topic. To do that requires a half-bridge electrical architecture not a simple high-side drive. You then have to deal with ground recirculation currents, so need a high current ground pin for the load as well as a high current supply pin. I'll have a think about that.

before you get too far ahead of yourself with this one, the only time I'm aware that this would be necessary is if you're trying to drive a DC motor either direction...on command...and change that direction whenever you want.

simple PWM control can be achieved with a high-side or low-side switch for fan speed and fuel pump regulation.

Owen Thomas

07-24-2013, 10:14 AM

Our team makes our own 'power distribution module', which is actually just a bunch of wires, relays, and fuses contained in one central location. It is very, very simple and it is easy to diagnose something like a blown fuse. It does not include the starter control (straight from the battery), it is not programmable in any way, and we like it. As a relatively low-tech team with somewhere between 0 and 1 EEs on the team at any given time, I can't honestly say anything with higher functionality would really be used.

The point is, if you're trying to cater to a mass market of mostly mechies, the simpler the better. As an electrically illiterate person with a good rudimentary understanding of basic electron-wrangling principles (read: more than average but less than adequate), I can confirm that I did not understand roughly 70% of Johns post above. I can also confirm that any item I purchase with adjustability I will assume should be adjusted to try and make it "better". Anything that can be tinkered with, will be tinkered with. This causes problems on both the supplier side and the consumer side, since the consumer will immediately blame the product when thier tinkering inevitably causes problems or poor performance.

All I'm saying is that you can't really be successful in making a product for both e-tards like myself and still have the higher level functionality desired by the more e-savvy people like jlangholz. This is merely my opinion, but like all designs you will probably have to make significant compromises. Personally, I'm in favour of the "set and forget" package, like Jim and Pete.

Oh, and our fan consistently blows 15A fuses. Pretty much everything else (fuel pump, ign, ECU) is run off of 5A or 10A.

jlangholzj

07-24-2013, 11:27 AM

Originally posted by Owen Thomas:
Our team makes our own 'power distribution module', which is actually just a bunch of wires, relays, and fuses contained in one central location. It is very, very simple and it is easy to diagnose something like a blown fuse. It does not include the starter control (straight from the battery), it is not programmable in any way, and we like it. As a relatively low-tech team with somewhere between 0 and 1 EEs on the team at any given time, I can't honestly say anything with higher functionality would really be used.

The point is, if you're trying to cater to a mass market of mostly mechies, the simpler the better. As an electrically illiterate person with a good rudimentary understanding of basic electron-wrangling principles (read: more than average but less than adequate), I can confirm that I did not understand roughly 70% of Johns post above. I can also confirm that any item I purchase with adjustability I will assume should be adjusted to try and make it "better". Anything that can be tinkered with, will be tinkered with. This causes problems on both the supplier side and the consumer side, since the consumer will immediately blame the product when thier tinkering inevitably causes problems or poor performance.

All I'm saying is that you can't really be successful in making a product for both e-tards like myself and still have the higher level functionality desired by the more e-savvy people like jlangholz. This is merely my opinion, but like all designs you will probably have to make significant compromises. Personally, I'm in favour of the "set and forget" package, like Jim and Pete.

Oh, and our fan consistently blows 15A fuses. Pretty much everything else (fuel pump, ign, ECU) is run off of 5A or 10A.

http://fsae.com/groupee_common/emoticons/icon_smile.gif whoops

on a side note, 15A for a fan...you guys must be running a 14" odd fan or something...

Ours is currently just a re-dux of the previous board that I had made. In years past it had been relays and fuses, which worked great and was a step in the right direction but it was HUGE! All that I did this year was add some FET's instead and make things smaller.

looks like this:

http://i260.photobucket.com/albums/ii35/jlangholzj/IMG_20130724_132842_zps39eb7879.jpg (http://s260.photobucket.com/user/jlangholzj/media/IMG_20130724_132842_zps39eb7879.jpg.html)

In addition, it would be pretty easy to make it "less e-savy capable"(?). A lot of the options I'm thinking about are done via computer...like what you'd do with an ECU. With the option of either having input pins or over CAN, you could set it up just like a "dumb board" that would also have the protection features like current monitoring.

DANGIT....if I don't have enough to do already....4 papers to write...and now you've got me going on this...ugh...i said i wasn't going to do this..

and don't worry owen, you can keep all that black magic suspension stuff.

murpia

07-24-2013, 12:42 PM

Originally posted by jlangholzj:
<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">Originally posted by murpia:
PWM control is an interesting topic. To do that requires a half-bridge electrical architecture not a simple high-side drive. You then have to deal with ground recirculation currents, so need a high current ground pin for the load as well as a high current supply pin. I'll have a think about that.

before you get too far ahead of yourself with this one, the only time I'm aware that this would be necessary is if you're trying to drive a DC motor either direction...on command...and change that direction whenever you want.

simple PWM control can be achieved with a high-side or low-side switch for fan speed and fuel pump regulation. </div></BLOCKQUOTE>
I think you're confusing a half-bridge with a full-bridge. A full-bridge is needed for direction control. The half-bridge I'm referring to is simply either a single high-side driver plus a diode, a single low-side driver plus a diode, or a pair of drivers. The key point is that there is route for the recirculation current during PWM either through the diode or the other driver.

Regards, Ian

jlangholzj

07-24-2013, 02:17 PM

Originally posted by murpia:
<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">Originally posted by jlangholzj:
<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">Originally posted by murpia:
PWM control is an interesting topic. To do that requires a half-bridge electrical architecture not a simple high-side drive. You then have to deal with ground recirculation currents, so need a high current ground pin for the load as well as a high current supply pin. I'll have a think about that.

before you get too far ahead of yourself with this one, the only time I'm aware that this would be necessary is if you're trying to drive a DC motor either direction...on command...and change that direction whenever you want.

simple PWM control can be achieved with a high-side or low-side switch for fan speed and fuel pump regulation. </div></BLOCKQUOTE>
I think you're confusing a half-bridge with a full-bridge. A full-bridge is needed for direction control. The half-bridge I'm referring to is simply either a single high-side driver plus a diode, a single low-side driver plus a diode, or a pair of drivers. The key point is that there is route for the recirculation current during PWM either through the diode or the other driver.

Regards, Ian </div></BLOCKQUOTE>

Caught me red-handed. Although the Half-H bridge that you mention (single diver plus a diode) is what I was talking about with the flyback protection diode, so we're on the same page here...just talking differently!

murpia

07-24-2013, 03:01 PM

Originally posted by jlangholzj:
<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">Originally posted by murpia:
<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">Originally posted by jlangholzj:
<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">Originally posted by murpia:
PWM control is an interesting topic. To do that requires a half-bridge electrical architecture not a simple high-side drive. You then have to deal with ground recirculation currents, so need a high current ground pin for the load as well as a high current supply pin. I'll have a think about that.

before you get too far ahead of yourself with this one, the only time I'm aware that this would be necessary is if you're trying to drive a DC motor either direction...on command...and change that direction whenever you want.

simple PWM control can be achieved with a high-side or low-side switch for fan speed and fuel pump regulation. </div></BLOCKQUOTE>
I think you're confusing a half-bridge with a full-bridge. A full-bridge is needed for direction control. The half-bridge I'm referring to is simply either a single high-side driver plus a diode, a single low-side driver plus a diode, or a pair of drivers. The key point is that there is route for the recirculation current during PWM either through the diode or the other driver.

Regards, Ian </div></BLOCKQUOTE>

Caught me red-handed. Although the Half-H bridge that you mention (single diver plus a diode) is what I was talking about with the flyback protection diode, so we're on the same page here...just talking differently! </div></BLOCKQUOTE>
A passive flyback diode will dissipate approx. 0.6W per Amp of driver current, compared to an active half-bridge which will be much less. Devices are available which incorporate current sensing too, hence my preference to use one of those.

Regards, Ian

Pete Marsh

07-24-2013, 07:21 PM

As a general query, how would you as users expect the CAN control to work?

Work? No idea. But I know what I would like it to do.
I was not initially thinking of something so small, now I see that, how about remote control of it.
It's the instant feedback and indication of the mechanical circuit breakers that is awesome, plus the ability to hold them in. I'm thinking a 6 station switch panel with CAN to the PDM. Momentary buttons software programmable to latch if desired, or simply be a breaker override and indicator if that's what you want. Each with multi coloured LED status indication for 'off' 'on open circuit warning' 'on OK' 'on over current' 'off tripped warning (flashing)'.
Also programmable ability to push override to a higher current setting and a master alarm output.
All mounted in a nice panel big enough to operate with gloved hands in a moving car. If you need more than six, just stack them up.

Can you do 30A on a single channel? If not then a 4 button panel might also be good so you can dual path 2 channels in the (same)PDM to get 2 big channels with two regular ones.

I'm not full bottle on the electronics, but is it feasible to put the tricky electronics and ECU coms in the remote panel, such that the bear bones PDM is really simple, just a tiny relay and fuse station, and the upgrade is to buy the switch panel for full function?
I could see your typical car using one switch panel and two PDMs for an all out system, so some functions are dumb analogue switched and some are CAN and fancy, and maybe one or two are doubled for extra capacity.

Pete

murpia

07-25-2013, 01:38 AM

Originally posted by Pete Marsh:
<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content"> As a general query, how would you as users expect the CAN control to work?

Work? No idea. But I know what I would like it to do.
I was not initially thinking of something so small, now I see that, how about remote control of it.
It's the instant feedback and indication of the mechanical circuit breakers that is awesome, plus the ability to hold them in. I'm thinking a 6 station switch panel with CAN to the PDM. Momentary buttons software programmable to latch if desired, or simply be a breaker override and indicator if that's what you want. Each with multi coloured LED status indication for 'off' 'on open circuit warning' 'on OK' 'on over current' 'off tripped warning (flashing)'.
Also programmable ability to push override to a higher current setting and a master alarm output.
All mounted in a nice panel big enough to operate with gloved hands in a moving car. If you need more than six, just stack them up.

Can you do 30A on a single channel? If not then a 4 button panel might also be good so you can dual path 2 channels in the (same)PDM to get 2 big channels with two regular ones.

I'm not full bottle on the electronics, but is it feasible to put the tricky electronics and ECU coms in the remote panel, such that the bear bones PDM is really simple, just a tiny relay and fuse station, and the upgrade is to buy the switch panel for full function?
I could see your typical car using one switch panel and two PDMs for an all out system, so some functions are dumb analogue switched and some are CAN and fancy, and maybe one or two are doubled for extra capacity.

Pete </div></BLOCKQUOTE>
My intent behind offering CAN control was to allow the PDM to be better integrated with other vehicle systems, by the users. I'd still be interested in opinions on how best to offer that? Decoding of ECU data? Dumb protocol?

If a straightforward 'remote control' is wanted, I think serial control is enough (and cheaper). The 'remote control' could use the same case & connector (with a different keyway), the same micro, etc. all helping keep costs down.

Regards, Ian

jlangholzj

07-25-2013, 02:37 PM

Originally posted by murpia:
My intent behind offering CAN control was to allow the PDM to be better integrated with other vehicle systems, by the users. I'd still be interested in opinions on how best to offer that? Decoding of ECU data? Dumb protocol?

If a straightforward 'remote control' is wanted, I think serial control is enough (and cheaper). The 'remote control' could use the same case & connector (with a different keyway), the same micro, etc. all helping keep costs down.

Regards, Ian

If you're not going to decode the CAN stream on the system, then a simple dumb analog in would be fine...aka just tying all of your ECU outputs into the box. Otherwise there's no real benefit right? Also using the CAN would require the user to be more savy (would have to enter in on/off values in the PDM control software)

I like the remote idea that pete has. As you may have (or may not have) seen I had LED's on our board for simple troubleshooting. Pulling these out to a remote board that has a "reset" switch persay wouldn't be too bad of an idea. Would basically be a software version of a circuit breaker. Also would feedback the current status of the output. Could use an RGB surfacemount LED to show different status; green = on, yellow = off, red = tripped...or something similar.

Pete Marsh

07-25-2013, 06:49 PM

One big advantage of a software controlled fuse/breaker system IMO is the ability to detect and display an open circuit fault. This is probably a more likely failure mode than a short circuit or overload, and stops the fun just the same.

Pete

murpia

09-11-2013, 09:19 AM

Hi everyone, time to revive this thread...

I have had time now to analyse the likely spec of a $200 PDM. Here's a preliminary spec, pinout & functional description for the stand-alone (no CAN) mode of such a device:

Connector: Deutch IPD DTF13-12P ( http://www.laddinc.com/component/k2/item/1181-dtf13-12pa-1112.html )

Pin Function:

1) 25A continuous on / off or PWM output, with 50A peak capability (thermal fuse emulation), open-circuit load diagnostic & retry strategy
2) 15A continuous on / off output, with 25A peak capability (thermal fuse emulation), open-circuit load diagnostic & retry strategy
3) 15A continuous on / off output, with 25A peak capability (thermal fuse emulation), open-circuit load diagnostic & retry strategy
4) 15A continuous on / off output, with 25A peak capability (thermal fuse emulation), open-circuit load diagnostic & retry strategy
5) 15A continuous on / off output, with 25A peak capability (thermal fuse emulation), open-circuit load diagnostic & retry strategy
6) Ground pin
7) control & status pin for output pin 1), switch to ground to turn on output, can be PWM'd to 1kHz
8) control & status pin for output pin 2), switch to ground to turn on output, 1s off / on debounce time
9) control & status pin for output pin 3), switch to ground to turn on output, 1s off / on debounce time
10) control & status pin for output pin 4), switch to ground to turn on output, 1s off / on debounce time
11) control & status pin for output pin 5), switch to ground to turn on output, 1s off / on debounce time
12) mode select pin for advanced features (topic for another post)

There will also be an M6 stud to provide the connection to the battery. I plan to support up to 30V battery voltage (i.e. 12 cell lead-acid or 8 cell LiFe).

The 1s on / off debounce time for outputs 2->5 is designed to prevent inadvertent PWMing of these outputs causing too many inductive transients. This does not mean the output takes 1s to respond to an on or off signal, but following an off signal it cannot be turned on again for 1s.

The open-circuit diagnostic is achieved by pulling the load to 5V via a resistor, so a few mA will flow in the load to detect connection to ground. This should not be enough to energise anything significant (fan, pump, etc.).

The thermal fuse emulation will allow an output to momentarily exceed it's rated current without tripping, although if the condition persists the output will trip. The more the current exceeds the rating, the faster the output will trip. The output will retry the load after 1s, if the load current is OK the output will stay on. If not, the output will retry the load after 2s, then 4s, 8s etc. to avoid damaging hardware up to a persistent retry time of 64s.

The diagnostic is indicated for an output by supplying a regulated 25mA to the corresponding control & status pin. Connecting a 25mA rated LED between the control & status pin and ground means the LED can be made to light whenever the control & status pin is not pulled to ground by the switch. On power-up, if the switch is open the LED will be lit steady to show a load has been detected OK, and flashed to show it has not. A different flash pattern will be used to indicate an over-current trip has taken place. Using the control & status pin this way simplifies the wiring, although has the drawback that the LED will go out whenever the switch is made regardless of load status. Without extra connector pins there's not much else that can be done, and pins add cost.

I plan to make the PDM much more configurable than this if CAN is used, or if the user plugs into the PDM and runs specific configuration software, but for an out-of-the-box setup how does the above sound?

Regards, Ian

EPMPaul

09-11-2013, 12:32 PM

murpia

09-11-2013, 02:48 PM

Only problem with DT series connector is that their load rating is 13A at max rated temperature. Obviously, I don't think your PDB is gonna hit 125 deg C but 70A through your ground pin might be a little ambitious.
With a single ground pin all loads have to return current through the wiring harness to the battery. This is normal PDM wiring practice. The single ground pin does need to flow PWM recirculation current, hence there is only one PWM output.

Also might be interesting to have a few always on fused outputs (for brake light and the sorts)
Just hard-wire the appropriate control pin to ground and you will get this.

Regards, Ian

EPMPaul

09-12-2013, 11:11 PM

Looks pretty sweet. For 200$ would definitely be worth it. Fair enough for the ground pin. On last year's loom I used one of these http://www.cooperindustries.com/content/public/en/bussmann/transportation/products/power_distribution/power_distribution_modules/series_15300_rtmr.html

While they're not pretty to look at and a little bit heavy they do a fair job of keeping things dry and safe. Cool thing about it the possibility of dual busbars so makes wiring super easy just run all your 12V and ground to that(other than the starter). Sort of why I was asking about the ground pin. While it seems to be done differently in your setup, makes more sense for newbs looking for ground issues/12V issues.

As for CAN control of the software, I worked for BRP(Bombardier Recreational Products ) a while back and they used a protocol called Keyword 2000 for reflashes and stuff like that. It was a royal pain in the ass when you had to query through it on a DAQ for some ATVsbut for programming it worked sort of like a Serial network. One CAN address was a TX and the other a RX. While that particular implementation was sucky because it was adapted from the ECU manufacturer so it had a necessity for a shit ton of funtionality(there were 5 bytes of protocol stuff and 3 bytes of actual data so not so efficient). The advantage was you just needed a CAN Sniffer to diagnose/reflash the ECUs. Only thing is you've got to make sure no other device uses a similar protocol or at least not on the same addresses otherwise it can create some serious havoc.

Basically, the the protocol bytes went along the lines of (Byte 1: destination(each node gets an ID), Byte 2: instruction(R/W), Byte 3 : Registry , Byte 4/5: Number of total CAN messages to transmit/ Current message in sequence .Rest of the bytes were the actual data

On the plus side, if you've got a bunch of devices running the same retarded protocol, you can control your devices with all the added reliability and noise proofing of a CAN bus. While on ATVs and such it's not such a problem because they're formated/reflashed in the shop, if you want to be able to reset breakers or whatever, you can create a control unit and not have to run a bunch of pins to your PDB. Reduction in cost and bulk of the overall system/ with some added complexity but anyways, just throwing it in there to say that there are already implementations of higher level protocols over the CAN bus in the industry. While direct application would probably proove a pain just coz these things tend to gob busload like crazy, you could probably make a simplified version.

On a more personal note, I looked into doin a PCB for power distribution this year and when looking at the potential gains, once you have a reliable loom, considering the above unit is 50$, the development time and potential flop of a mechie going anywhere near a PCB kind of means that horse got shot in the face for now due to other responsibilities on the team.

On that note back to work

murpia

09-13-2013, 04:59 PM

Just a quick description of my plans for the mode select pin:

In some setups there's a need to be able to kill power to the loads based off the driver's 'ignition' switch. For example most engine ECUs suggest you power the injectors, coils & fuel pump off a switched ignition source. With a PDM based setup this can be more problematic as the intent of a PDM is that it's hard-wired to the battery for maximum reliability.

The mode select pin can help here, by default it must be connected to battery to enable the module. This is known as 'ignition sense'. The ignition switch can then be a low-current type but still be used to switch high-current loads via the PDM.

Additionally I plan to multiplex 2 of the control pins with a CAN transceiver. CAN mode will allow the PDM to be controlled remotely and to report the load currents for datalogging. CAN mode will be entered by wiring a number of the remaining 3 control pins to mode select. With 3 pins that allows for 7 configuration selections to be made. The exact details of these is TBD but will probably include a selection of CAN IDs so that multiple PDMs can be combined on a single bus.

I've selected most of the components now including the micro. Next steps are to build some prototype circuits for test. If they work OK an initial batch of PCBs will follow.

Regards, Ian

Pete Marsh

09-16-2013, 06:36 AM

murpia

09-16-2013, 07:43 AM

I'm not sure I get the fault monitering function. It only works when the function is switched off? Does that make any sense? Can you do more if the LEDs are just on the main unit? Or maybe a second tiny connector (rc battery balance lead maybe) just for the monitering lights?

I like the "ign sense" idea.

Pete
Fault monitoring works all the time, it's just that the display of the error condition can only happen when the switch is off (without adding pins).

There are a few reasons for this:

Cost - adding another connector just for lights increases the PCB, case & micro costs
Wiring practicality - the proposed warning light design needs no additional wiring to / from the module, just an LED wired locally across the switch on the dash panel
Complexity - as soon as you start complicating the system you are going away from the minimalist, low cost solution everyone was interested in

On the complexity issue, that's what CAN mode is for, you can have all sorts of dash alarms, error logging, diagnostics with CAN mode. But CAN only will put off a lot of toe-in-the-water type customers so I have designed the stand-alone mode to cover most possibilities.

I would propose to eventually design a CAN mode supervisory controller, which would mount on the dashboard and have a display plus connections for lights & switches. The hardware for that is simple, easier actually than the PDM itself. But the software, both embedded & PC configurator side, is a lot of work if you want to provide a slick, easy to use UI.

Regards, Ian

murpia

09-17-2013, 07:12 AM

Just a query on the main power connection (needs to supply up to 85A continuous).

I was planning an M6 stud like the MoTeC units. But another option is a connector like this:

http://www.laddinc.com/our-products/online-catalog/item/1210-dthd06-1-8s-1141.html

Or this:

http://www.andersonpower.com/files.php?file=DS-PP75(10).pdf

Opinions?

My main concern with a stud is if it comes loose you may get sparking / arcing = a fire. Also torque reaction needs to be contained when the user is doing it up.

My main concern with a connector is cost (including tooling for the user), followed by security - how likely is it to become accidentally disconnected. The Deutsch connector can be secured with a cable tie, so may be the best best?

Regards, Ian

Kevin Hayward

09-17-2013, 10:53 PM

Ian,

I wouldn't be worried with any of the three methods of connection. Teams are generally used to working with stud connections on batteries and putting in a locking method is something teams do on nearly all bolted connections (or should do). I think you will want very clear installation instructions for the customer. If you use studs you can suggest some locking methods in those instructions.

Talking to our guys I think a stud would be better as it will proabbly take the least space. The smaller this box is the more we want one. Although from what you have already mentioned we will be after a couple anyway.

Kev

murpia

09-18-2013, 07:16 AM

Thanks for the comments, I think I will go with the M6 stud initially.

I have an opportunity next week to do some PCB layout. After that I should be able to indicate dimensions accurately, but my target is 100mm x 50mm plan view, 40mm high... I may adjust that a little depending on the availability of a suitable case extrusion profile.

Regards, Ian

jlangholzj

09-18-2013, 01:49 PM

Thanks for the comments, I think I will go with the M6 stud initially.
.....the availability of a suitable case extrusion profile.

Regards, Ian

Why not make one? You aren't going to need SLS rp to do a case. The much cheaper option of SLA would work perfectly fine and allow you to mold to the 6-pin connector as well as the ground stud.

Additionally, if you would like to get away from the stud multiple pins can be connected to the ground plane of your PCB (you are having a ground plane...right? with separated analog and digital planes??). I did this with our design this last year to much success. I also used multiple output pins for different output functions (ie, if its drawing more current use two pins). Internal temperatures seemed to be very happy and we haven't had a failure on it yet.

Also for those of you worried about cost effectiveness:
there's no doubt that having a simple, clean, reliable loom is priority #1!!! but we get our PCB's spun for free via sponsor and electrical components for free via another sponsor. I'd say it's pretty cost effective! Once you get a good design tested, make TWO. That way if something really goofy happens, a quick plug and unplug and you're fixed on raceday without any downtime.