Jump to content

Separate Power board, tech spec controller


Recommended Posts

Hi all!

I have search a lot around this forum and other pages but still haven't found enough detail about the controller card of the Gotway EUC's. In particular the Msuper 3 since I'm currently working on a separate power-board with either 6 sturdier MOSFET's  or 12 smaller as the original. Since most controller cards share the same basic setup I suspect that we could gather "all knowledge" in the same thread.

One reason to go for 12 MOSFET's is the fact that there are components that them self may cope with currents above 360A but the connector leads stop at around 160-180A. So the first data I'm asking the forum for is

Q1: Has anyone done actual measurements on the currents fed thru the motor of a Msuper 3? Preferably also an oscilloscope image...

 

I have read a lot of documentation and can really recommend the Application Note AN-941 from where I quote the summary

It is advisable to follow these general guidelines should be followed when paralleling MOSFETs:

  • Use individual gate resistors to eliminate the risk of parasitic oscillation. Ensure that paralleled devices have a tight thermal coupling.
  • Equalize common source inductance and reduce it to a value that does not greatly impact the total switching losses at the frequency of operation.
  • Reduce stray inductance to values that give acceptable overshoots at the maximum operating current. Ensure the gate of the MOSFET is looking into a stiff (voltage) source with as little impedance as practical.
  • Zener diodes in gate drive circuits may cause oscillations. When needed, they should be placed on the driver side of the gate decoupling resistor(s).
  • Capacitors in gate drive circuits slow down switching, thereby increasing the switching unbalance between devices and may cause oscillations.
  • Stray components are minimized by a tight layout and equalized by symmetrical position of components and routing of connections.

 

I myself have not seen the 6 MOSFET-version of Gotways circuit boards but as I have understood the "fix" is to only solder another set on top, there's no other change of components, the impedance from the driver has been left unchanged. Also, as most of you have seen, there's no individual gate resistor, only a true parallel connection with the extra MOSFET simply mounted in the same PCB-hole. This may not be so serious depending on what frequency the PWM is operating at. Which leads med over to the next question

Q2: Has anyone measured the PWM frequency of the Gotway controller board?

 

But in general I sort of have a feeling that Gotway had a larger stock of controller cards and the add-on MOSFET is a quick fix. This is one reason they are somewhat reluctant to give hardware support and spare parts, many small re-sellers have tried and failed. It's more a question of selling new EUC's than fixing the ones that break.

The plan now is, as I said initially, to build a separate power board. I de-soldered all MOSFET's along with the two 1200μF capacitors. Also all three motor-wires but realised that two of them are measured for current (the ACS709 circuit). Since there's no problem in that area at all I'll feed the current to "MotoA" and "MotoC" back to the controller PCB, keep it simple. Also, the main battery connector will be moved to the Power board, all the high currents will stay there. I will use only small signal wire connection to drive the controller card. The Power board will be mounted on the heat sink and a bigger heat sink will help balancing the heat in between the MOSFET's. Won't fit that big tunnel-thingy with a fan attached but some fan with thermostat will be used. One major drawback is that the electronics is cooled down partly by "air-speed" i.e. going slow on a hot day puts cooling to a test. The stripped controller PCB will fit nicely under the siren, it needs no additional cooling of any sort.

If I get no response on the question above, I'll do some measurements o my own, just being lazy here ;0)

But, please help out and collect as much electronic data as possible in this thread! Or link to other threads, I'll try to compile all data here if it seems to be spread out...

heatsink.png

PCB_front.jpg

PCB_rear.jpg

PCB_rear_burn.jpg

Link to comment
Share on other sites

PWM is around 7-9khz. Most of the technical details are in the custom firmware thread in the mods forum. Currents in the more powerful wheels max out at around 50A-100A in phase, but only in short peaks and lower duty cycles where the inverter converts the battery energy into a lower voltage and higher current they never draw that much from the battery. Realistically they can deal with maybe 20A or 30A continuously, many have 40A fuses on the connection to the battery.

I gave up on using any of the existing boards, they have too many flaws. I have been working on a custom board, its almost done I just still need someone to help with the balancing code. This is my most recent board, pretty much finished but there was one mistake on the PCB so have to fix it and order another.

 

DSC_0278.JPG

Link to comment
Share on other sites

1 minute ago, frenchie said:

Too bad that's 38 pages about everything but the Gotway. But I'll search it for data...

They are basically all the same board on every EUC 500W and above. No company his really invested any effort in control boards the last year or so.

Link to comment
Share on other sites

With a Wheatstone bridge I now have measured the resistance in the motor windings, thru each connection = blue-green, blue-yellow, green-yellow. Value was, as expected, the same on all connections i.e. two coils in series = 139 milli-ohm = 69.3 milli-ohm per coil.

The original MOSFET IRFB3207 is suposed to have maximum RDS(on) = 4.5 milli-ohm and therefore chews up some Power.

P = I2 * R

Total R = 69.3 + 69.3 + 4.5 = 143 milli-ohm.

The original MOSFET IRFB3207 is supposed to deliver 180A which thru the formula above leaves us 2247W per coil but, unfortunately also 146W in the component itself. This is of course "worst case" values. The truth is that the original MOSFET put in a TO-220 casing only can deliver 75A due to thin connectors. That is partly the reason Gotway doubled up by soldering a twin MOSFET in top of the original.

The value on Rjunction-case is maximum 0.45 C/W for the IRFB3207. If it has been properly mounted on a heat sink that is!! This gives us 0.45 * 146 = 65.6 degrees C in worst case. In reality we might have a current As a base addition to the heat needed to get rid of that's a little too much.

Kept in mind that the controller PWM-signal wont send out any different signal than before (open MOSFET is an open MOSFET, it will not send more Amps to the motor than the original setup, it just copes better with the current), the total current available is still set by the voltage/total resistance with a theoretical burst maximum at 67.2V/143milli-ohm = 469A (actually slightly higher since the max current should be calculated with the minimum total resistance, typical RDS(on) 3.6 milli-ohm = 472A). The average current is what is driving the wheel! IRFB3207 can cope with maximum burst current = 720A. The original design with only 6 MOSFET's seems to be correct. But the heat takes the better of them!! The heat sink is badly designed and much to thin where it counts. The average current might be as stated above by @lizardmech, a mere 20-30A.

 

However;

The MOSFET IRFP4368PBF has only 0.29 C/W and whats even better, RDS(on) = 1.85 milli-ohm! It also withstands a whopping 350A and as much as 195A in the connectors (TO-247AC is beefier than TO-220). Just to compare the worst case figure from the original MOSFET I use 180A again

P = I2 * R

Total R = 69.3 + 69.3 + 1.85 = 141 milli-ohm.

Pcoil is still 2247W

But the PMOSFET = 60W. And with a lower Rjunction-case at 0.29we only raise 17.4 degrees C, not 65.6!

 

Replacing to IRFP4368PBF will be easy to cool down, sturdier current-wise and absolutely not demanding twin-MOSFET. It'll be more than enough with 6 units. That way all the other problems Gotway has completely disregarded (or just failed to calculate) when the emergency soldered twin-MOSFET's, totally vanish. For instance, the biggest no-no is the 90-degree bend immediately right next to the casing, that's not OK! Also, the twin doesn't share the same heat sink profile. That gives you different temperature per individual MOSFET and therefore a large risk of switch oscillation. The MOSFET twins running forward thru "MotoB" also just happen to be in the middle heat-wise. This also builds up a much to high temperature in general.

Guess where I will mount my new MOSFET's ;0)

Link to comment
Share on other sites

You have to consider switching losses as well. Comparing the two infineon fets and the fairchild fet I'm using, the  IRFP4368PBF has low RDSon, however input capacitance is twice the original and rise/fall times are dramatically slower, it's also limited to 190A due to wire bonding in the package. The 12 fet board would suffer similar issues due to capacitance being doubled. In comparison the fairchild SMD fets have switching characteristics close to  IRFB3207 but the low 1mohm RDSon as well as a package rated for 300A.

The IRFP2368 will also spend more time in the miller plateau where resistance will be 3x higher. 

Link to comment
Share on other sites

On 1/11/2017 at 4:45 PM, frenchie said:

You are absolutely right. But at as low switching frequencies as this board use it shouldn't be that big issue?

 

What Fairchild SMD?

I think the issue is at lower currents the losses quickly add up. Even if it's just 2watts per fet you then have 12 watts heating the heatsink at all times, then when you do put a heavy load on it the fets are already at 60c+ and have a higher RDSon as a result.

These are the mosfets I have been using.

mouser.com/ProductDetail/Fairchild-Semiconductor/FDBL0150N80/

Link to comment
Share on other sites

  • 3 months later...

Funny, I've upvoted the first couple of posts back in Jan 7th, but have no recollection of ever reading this thread... Must have been drunk or something :P

But anyway, I'd be really interested in hearing how this turned out. Another thing I was wondering was that if the original gate-drivers don't pack enough "punch" to drive the higher gate charge-mosfets fast enough, do you think it's possible to buffer the gate-signals and use a separate card with it's own driver(s) and mosfets? What problems are likely to be faced, interference, too much delay between the original signal and the fets starting to conduct..?

Link to comment
Share on other sites

  • 7 months later...

Did this one die out, moved?

Not that I can follow 100% but interesting nevertheless and just send my ACM to shop today for inspection so..

 

Also thanks for the images, another image I see that have got no solder/splatter connecting the Green motor wire (C) to the 2mΩ just next to it like mine has, hard to accept when no other boards have it but not necessarily the reason for my sudden loss of balance either the whole board looks like a 5 year old played with fathers soldering gun and somehow manage not to get burned in the process and/or exposed?

The board though, ohh well. The pig fest reference in general is worthy, wish I had never seen that.

Link to comment
Share on other sites

Archived

This topic is now archived and is closed to further replies.

×
×
  • Create New...