DIY PSU Build for Flashing

[HDDMark]

I think my lack of electrical circuit knowledge is catching up with me.

My thought was that, even though the PSU is set at a constant 13.6 volts and I never saw a voltage on the car terminals even approaching that, the battery in the car wasn't seeing 13.6 volts either. If that was the case, we could increase the voltage at the PSU to up the current only so long as the voltage at the terminals never exceeded 14 volts. If you could regulate the PSU to sense the voltage at the car's terminals and keep it's voltage at a safe level at the terminals (<14 volts) we'd be in good shape.

Based on what you guys are saying, it doesn't matter what the voltage at the car's terminals is, the car's battery will still see something close to the PSUs output voltage and cut the circuit to protect the battery.

It seems that this PSU needs to be able to provide a higher current at a lower voltage difference to be effective for flashing purposes. I guess that's where multiple PSUs in parallel would help. Of course if using the different sense pins on the PSU (as you guys discussed earlier) allows it to put out more current at a given voltage difference that would be awesome.

[terahertz]

Originally Posted by HDDMark

It seems that this PSU needs to be able to provide a higher current at a lower voltage difference to be effective for flashing purposes. I guess that's where multiple PSUs in parallel would help.

I would agree with this, somewhat. Multiple PSUs in parallel would only help if the single PSU cannot provide enough amps. Based on the specs of the PSU, and based on the current draw that was measured by ferrarif1's charger, a single PSU should be more than enough.

I feel that perhaps your connection to the car has too high of resistance or is not making good enough contact. This would make it "harder" for the car to draw power from the PSU.

I am building a single PSU for jwzimm, so hopefully next week we can hear how it works for him.

[aboulfad]

@HDDMark , here's the FUB on our batteries: https://www.dropbox.com/s/yr8ohfcn71...ttery.pdf?dl=0 : the maximum voltage not to exceed is 14.4V otherwise the disconnect switch disconnects internally and the battery would show 0V! (Applying a power pack or a charger capable of detecting low or no voltage will reactivate switch)

Also, we haven't discussed or know the operation and behaviour of that PSU when faced with current surges... I think not knowing any of this, I would use two in parallel!

[jwzimm]

I will be happy to report back results but I don't have any flashing activities in my near future. My firmware was updated last week at the dealer to resolve an issue with the Kombi wigging out so I am on the latest and greatest right now. I mainly want the power supply so I can code without running the battery down or running the engine.

[aboulfad]

Originally Posted by jwzimm

... I mainly want the power supply so I can code without running the battery down or running the engine.

I think coding will consume way less current than flashing, and you are guaranteed during coding not to have water pumps running but still interested in any feedback you have.

[jwzimm]

Originally Posted by aboulfad

I think coding will consume way less current than flashing, and you are guaranteed during coding not to have water pumps running but still interested in any feedback you have.

Yep. That is one of the reasons I opted for a single unit PSU.

[jwzimm]

So I am looking over the rcforums thread and I think the ground isolation is not needed for what we are doing. The only reason to go inside the PSU and isolate the ground from the chassis of the PSU is if you are going to be running two PSU's in series to produce 24V. In that case you would only want to isolate the ground on the 2nd PSU since it would be getting fed the +12V into its ground port which would cause the entire chassis of the 2nd unit to go "hot".

For what we are doing there is no reason to isolate the ground internally. In that case there is no reason to open up the PSU at all. You would just need to add on the jumpers for the settings and solder on the output leads to connect to the car.

[terahertz]

Originally Posted by jwzimm

So I am looking over the rcforums thread and I think the ground isolation is not needed for what we are doing. The only reason to go inside the PSU and isolate the ground from the chassis of the PSU is if you are going to be running two PSU's in series to produce 24V. In that case you would only want to isolate the ground on the 2nd PSU since it would be getting fed the +12V into its ground port which would cause the entire chassis of the 2nd unit to go "hot".

For what we are doing there is no reason to isolate the ground internally. In that case there is no reason to open up the PSU at all. You would just need to add on the jumpers for the settings and solder on the output leads to connect to the car.

Can anybody else confirm? Maybe someone who is an electrical engineer? I already isolated the ground on both of mine so I can't go back and test. If this is true, the DIY would be very simple!

[jwzimm]

Originally Posted by terahertz

Can anybody else confirm? Maybe someone who is an electrical engineer? I already isolated the ground on both of mine so I can't go back and test. If this is true, the DIY would be very simple!

The author of the article you linked actually confirms this. Check the very end of his first post.

you can parallel outputs to double your current if necessary. you'll have to connect pin 11 on each pair of voltage matching power supplies so they share the load. so say, you wanted 12V @ 1150 watts you'd just jumper pin 11 on 2 12V power supplies with their ground still in tact and connect positive to positive and negative to negative on the outputs.

[terahertz]

Originally Posted by jwzimm

The author of the article you linked actually confirms this. Check the very end of his first post.

Hmm. Okay. I will edit the DIY!

[aboulfad]

Originally Posted by terahertz

Hmm. Okay. I will edit the DIY!

No don't just yet... The topic is so complex and yes I am an EE and knows zilch about floating vs grounded PSU. We need an electrician for that my 2 cents, one primary reason for floating the DC on that PSU, is that if there's a short or other internal malfunction, so the AC current doesn't flow to the case and if few of them connect, boom! In essence, you want the low voltage side to be isolated from the high voltage side. The DC was connected to "chassis ground", the mod disconnects the DC from the chassis and is in a sense "floating". So is this fact less important when wiring in parallel?

I did lots of readings and in the end I had a big headache and decided to float the DC as many versatile expensive PSUs have DC floated vs grounded... Btw don't take my word for it, do the same searches I did by reading white papers from companies that specialize in PSUs not just rcgroup, ...

[terahertz]

Okay I reverted it but added a blurb about it until it can be confirmed.

[aboulfad]

Originally Posted by terahertz

Okay I reverted it but added a blurb about it until it can be confirmed.

Did jwzimm ask not to do it so you charge him less ? I feel it's safer to "isolate" the DC voltage from the chassis ground that the AC ground is also connected to. But to his credit, most posts mentione DC ground isolation only in series setup.

The question to ask, is it less safe in our parallel configuration to do so? I will ask a friend who's a real EE and he's amazing with circuits to try to get another opinion. (My brain is rusted from all of those management years)

BTW, I think this is the big thread over at RC http://www.rcgroups.com/forums/showthread.php?t=1581061

[jwzimm]

Originally Posted by aboulfad

Did jwzimm ask not to do it so you charge him less ? I feel it's safer to "isolate" the DC voltage from the chassis ground that the AC ground is also connected to. But to his credit, most posts mentione DC ground isolation only in series setup.

The question to ask, is it less safe in our parallel configuration to do so? I will ask a friend who's a real EE and he's amazing with circuits to try to get another opinion. (My brain is rusted from all of those management years)

BTW, I think this is the big thread over at RC http://www.rcgroups.com/forums/showthread.php?t=1581061

My understanding (being only a Mechanical Engineer) is that the only reason to isolate the ground from the chassis is to allow you to run a pair of PSU's in series without killing yourself. If you did not isolate the second PSU then it's chassis would be at 12V from the first PSU. If you were to touch that you would be seriously shocked or killed.

What we are doing is actually using the PSU to do what it is designed to do so there is no reason to isolate it.

I didn't identify this until after we agreed on a price and he built it. He has already done the isolation work. If I were to do it again I would request it to not be isolated but I wouldn't expect it to be cheaper.

As far as safety of running it isolated, I am curious to get the opinion of a EE as well. I can't really think of why it is not safe. The bottom line is that I can not see why you would isolate for our purposes.

[HDDMark]

Originally Posted by jwzimm

You are correct in regards to using the sense control to adjust for voltage drop as it may drive the power supply into overvoltage protection. I do think, however, that the basic voltage output control would be better by connecting pins 7 and 9. Here is why:

Pins 7 and 9 are the actual + and - Sense lines. These are tied into the 12V rails for their reference. The concern I have is that the power supply has no way to "See" the voltage it is putting out. By connecting pins 3 and 9 you are feeding a regulated 5V signal (plus Pot resistance) into the voltage regulator of the power supply. In this case the power supply will allow the voltage to fall as current draw increases since the 5V rail will not fall off. If, instead, we connected pins 9 and 7 together (through a different Pot) they would be referenced to the 12V rail which would allow for proper feedback and regulation of the 12V system. It would still not go above out setpoint of 13.6V and, therefore, not trigger protection but it would allow the supply to maintain a more constant voltage I think.

Thoughts?

JW,

Regarding the bolded Blue text, I'm unclear how this setup would work. Are you saying that I would no longer need the pot between pins 3 and 9 and replace it with sense wires to the car's terminals with another pot tuned to trick the PSU into upping the voltage? Would this setup then cause the PSU to put out more current at the same voltage drop or just up the voltage until it shuts off because it ups the voltage until it goes into overvoltage protection?

I guess my concern is that for flashing purposes this PSU needs to put out 20-30 amps at a voltage drop of less than about 0.2 volts (13.6 at PSU and 13.4 at car) or so and based on my observations with about a 0.5 volt drop during flashing it only puts out 10+ amps. I know that I can help my situation with larger and shorter leads but the weird thing is, you would think a PSU for a server would be pretty sensitive to voltage drops and put out more current with a small reduction in voltage.

I'm a civil engineer so I'm a bit behind you MEs and EEs. I think I need to google some of this stuff.

[aboulfad]

@jwzimm , I was only being silly about the $$$. Btw, one comment about the intended usage of the PSU, they are server rack mounted PSUs, and as such if they needed to be handled, you yank them out using the black plastic handle so small risk of human injury.

As to the DC ground isolation, the risk and danger is the flow of AC current to the chassis of the 2nd PSU in series, if there's a short or malfunction, given AC and DC share the chassis ground which is connected to earth ground.

Now what would happen if one would short out in the server configuration? and given they also run in parallel using that CSP, one could infer that there is no need to modify them in parallel configuration just like in the server application.

P.S. How many engineers does it take to build a PSU : (so far we have an EE, ME, CE)

[jwzimm]

Originally Posted by HDDMark

JW,

Regarding the bolded Blue text, I'm unclear how this setup would work. Are you saying that I would no longer need the pot between pins 3 and 9 and replace it with sense wires to the car's terminals with another pot tuned to trick the PSU into upping the voltage? Would this setup then cause the PSU to put out more current at the same voltage drop or just up the voltage until it shuts off because it ups the voltage until it goes into overvoltage protection?

I guess my concern is that for flashing purposes this PSU needs to put out 20-30 amps at a voltage drop of less than about 0.2 volts (13.6 at PSU and 13.4 at car) or so and based on my observations with about a 0.5 volt drop during flashing it only puts out 10+ amps. I know that I can help my situation with larger and shorter leads but the weird thing is, you would think a PSU for a server would be pretty sensitive to voltage drops and put out more current with a small reduction in voltage.

I'm a civil engineer so I'm a bit behind you MEs and EEs. I think I need to google some of this stuff.

Connecting pins 7 and 9 through an approximately 2.1kOhm resistance will result in about 13.6V output. You would not have anything connected to pin 3. What is being done by connecting Pin 3 to Pin 9 through a lower value resistor is feeding a set voltage back to the voltage control circuit. Pin 3 outputs a regulated 5VDC. The issue with doing it that way (3 to 9) is that the PSU is not able to sense when the output voltage drops due to load. If, instead, we use pins 7 and 9 the reference voltage being sensed is from the 12V rail that is being used to output. This means that the PSU is able to see and react to voltage drops.

This PSU is extremely sensitive to voltage drops but you have to feed it an accurate signal in order for it to compensate.

[jwzimm]

Originally Posted by aboulfad

@jwzimm , I was only being silly about the $$$. Btw, one comment about the intended usage of the PSU, they are server rack mounted PSUs, and as such if they needed to be handled, you yank them out using the black plastic handle so small risk of human injury.

As to the DC ground isolation, the risk and danger is the flow of AC current to the chassis of the 2nd PSU in series, if there's a short or malfunction, given AC and DC share the chassis ground which is connected to earth ground.

Now what would happen if one would short out in the server configuration? and given they also run in parallel using that CSP, one could infer that there is no need to modify them in parallel configuration just like in the server application.

P.S. How many engineers does it take to build a PSU : (so far we have an EE, ME, CE)

As always, more engineers leads to more confusion.

[aboulfad]

Ok one final point and we really "beat the horse dead", this thread is of the same opinion on isolating the DC ground regardless of the configuration: http://www.rc-help.com/content.php?4...POWERFUL%21%21

So terahertz, since this is your thread, you choose

[jwzimm]

Originally Posted by aboulfad

Ok one final point and we really "beat the horse dead", this thread is of the same opinion on isolating the DC ground regardless of the configuration: http://www.rc-help.com/content.php?4...POWERFUL%21%21

So terahertz, since this is your thread, you choose

Again, he is constructing a 24VDC power supply by running them in series. That does, indeed, require that one of the supplies has its DC ground isolated from the chassis. Also, note, he is only isolating the DC ground on one of the PSU's, not both.

His main concern is that a potential difference could be generated between the DC grounds of the two PSU's. This would result in an arc. That cannot happen with how we are using them. We are hardwiring the DC ground leads together thereby eliminating an possible potential difference.

[aboulfad]

jwzimm , I think you read diagonally!

"DC Ground - AC ground Isolation:

Why is it needed: DC ground on these power supplies is tied to the AC Ground. (try checking continuity between DC ground and AC ground pin or the body) So if you are running two simultaneously .. in Parallel or in Series, they cant touch.. Cause the DC potential being different on both, the metal bodies ( AC Ground) will be on two different potentials and if they touch … PZZZZZZZZZZT!! Sparks fly! "

Isolating DC ground is safer and more flexible, I have nothing else to contribute to this

[aboulfad]

Ok, finally made another custom shorter AWG8 cable, PSU Output set to 13.45V, and only drops to 13.40V with ignition ON, fan/AC max, low beams and halos, radio full blast

It's too late to hook up the 2.5Kohm pots so I'll try that in the weekend.