Cheezeball DC Load: DL24P: Pump, or Dump ???

[HerbTarlek]

Acrylic is flammable isn't it ? 


Burn baby burn....... 

Anything is flammable... if you have the right attitude.

[beanflying]

So Track sizing. Really rubbery here as it is pseudo guesstimate based 

As best as I can work out the board 'appears' to be 1 oz copper and not what really be a 2 oz job. Minimum trace widths (circa 8mm) and maximal lengths of run (80mm) when calculated based on this rubberiness indicate that going beyond 20A is going to be close to giving issues. Might be worth stripping the solder mask back on the power traces and applying some solder if you want to push the current toward or beyond 15-20A to keep the losses down.

Quick calculator used - https://www.7pcb.com/trace-width-calculator.php

Next bit is the Fet and power circuit and see where it might need a few additions based on a little TEA discussion a few weeks ago.

[fubgumfaw]

snip

The shunts are 5w, R010 or 10 mOhms. I have tons of shunts because of upgrading BMS's. I've touched my shunts under load and they do warm up a little. Replacing them with bigger ones is probably a good idea. They are probably OK for 20 amps as is.

I have to admit...china makes some cheap stuff and often times that means cutting corners...and yet I keep going back for more china junk...LOL!

I did see what you had put down but even allowing for 5W being a typo your 20A 'ok' is flawed so I ran the numbers and looked and some basic spec sheets.

I posted later that I used 10w shunts and other things I did to mine. I guess you missed those other posts. And yeah...this was a flat guess.
A little breaker in the current path will hold up pretty well for as often as it will need to break the circuit. No idea if it has an arc breaker in it. Low voltage, it won't matter much, but get the voltage up a ways and the arcing can be significant without one. Still...wish I had one on mine before I killed the gate drive on it.

https://www.ebay.com/itm/Carling-30A-Push-to-Reset-Button-Breaker-CLB-303-11ANN-3BA-125v-250v-AC-35v-DC/333592636271?hash=item4dababab6f:g:XlIAAOSw2pFesx5u

[beanflying]

The only reference I saw to the standard shunts was as I quoted and as there was no maths behind it so I did it. On mine I want to avoid to much work as I am happy to accept the Power limits as is and if I want to push it I will grab another later to really hack.

Still working out what is an easy fix in the stock space. Even sticking to 0805's is a bit tight for width so thinking I will tack one of these onto a Mouser order I need to make. My stash only has 0.001 and 0.0005's in it at present

https://au.mouser.com/datasheet/2/427/wslp2726-1762042.pdf

https://au.mouser.com/ProductDetail/Vishay-Dale/WSLP27265L000FEA?qs=sGAEpiMZZMtlubZbdhIBIO8cTNhIMphwPloVxIAYfvQ%3D



@brumby let me know if you want one tacked onto my order and I will throw it in the post when it arrives.

[fubgumfaw]

going beyond 20A is going to be close to giving issues. Might be worth stripping the solder mask back on the power traces

I did this too. Then I cut strips from 26 gauge copper sheet and bent it to fit the bare areas on the traces. Copper is so ductile that even a flat section stretches and compresses nicely to make bends and angles. It's hard to see it, but there is added copper under the mosfet and down to the mosfet leg. I'm pretty sure I'll not exceed the current handling of those traces! A+ and A- have copper soldered right to the pin on the  screw terminals.

[fubgumfaw]

The only reference I saw to the standard shunts was as I quoted and as there was no maths behind it so I did it. On mine I want to avoid to much work as I am happy to accept the Power limits as is and if I want to push it I will grab another later to really hack.

Still working out what is an easy fix in the stock space. Even sticking to 0805's is a bit tight for width so thinking I will tack one of these onto a Mouser order I need to make. My stash only has 0.001 and 0.0005's in it at present

I used 2 of these. They fit the space pretty well.
https://www.mouser.com/ProductDetail/71-WSHP2818R0100FEA

[fubgumfaw]

What I was doing for about a minute. 26 amps! Ha! Sorry for the glare on the LCD. Then the HY5110 mosfet died and so did the gate driver pin in the CPU. It still powers up, but getting the mosfet turned on no longer happens. Checking the gate for a signal shows nothing now.

[fubgumfaw]

In another test before reinforcing the copper traces and I guess running at 10 amps, this trace burned through. The 2 factory shunts burnt too. This was the first trace to get stripped off and reinforced.

[mawyatt]

Don't have this Load device but been down the Cheezeball equipment road a few times So take this lightly

If the Power MOS device gate is being taken out by over voltage oscillations as mentioned, if so, this is likely due to a feedback issue caused by the Power device "Miller" effect. Using higher power devices will also increase the Miller effect capacitance, so some form of compensation may be required. Placing a lead capacitor around the series gate drive resistor should help, also if a series resistor with the gate driver input, use a lead cap here. Protecting the power device with a Zener is a good idea, and protecting the CPU pin with a series resistor (or larger value if already used) may help, use a lead cap around this resistor as mentioned tho, otherwise you likely will have a big high powered oscillator!!

Anyway, good luck with getting this Cheezeball loads working for reasonable bench use.

Best,

[fubgumfaw]

Don't have this Load device but been down the Cheezeball equipment road a few times So take this lightly

If the Power MOS device gate is being taken out by over voltage oscillations as mentioned, if so, this is likely due to a feedback issue caused by the Power device "Miller" effect. Using higher power devices will also increase the Miller effect capacitance, so some form of compensation may be required. Placing a lead capacitor around the series gate drive resistor should help, also if a series resistor with the gate driver input, use a lead cap here. Protecting the power device with a Zener is a good idea, and protecting the CPU pin with a series resistor (or larger value if already used) may help, use a lead cap around this resistor as mentioned tho, otherwise you likely will have a big high powered oscillator!!

Anyway, good luck with getting this Cheezeball loads working for reasonable bench use.

Best,

 Sorry about the glossy look. I coated all the small components in conformal.

We add a 15v zener at the mosfet. This can be seen in the second image. The below image is the gate driver complete with a resistor in series to the large mosfets gate. That lower left leg has a resistor on it and then the trace goes directly over to the large mosfet gate.


 
To the lower right of the main mosfet is a thermistor. This is pretty much the entire driver circuit for the mosfet gate. Far left is the same components shown in close-up in the above image. The 3 legged component on a small tab with 2 SMD caps is some kind of v-reg. My zener is disconnected ATM since I'm trying to trouble shoot why my gate circuit isn't working. Perhaps a 10K resistor from gate to source in parallel with the zener will help? Maybe the placement of the zener is sub optimal?

[mawyatt]

I suspect the Zener is fine. What value is the gate series resistor? You could place a small cap on top as the lead cap, without a schematic and just guessing I would make the RC time constant ~100us. Also is there a series resistor to the CPU output? If so might consider placing a lead capacitor over this resistor as well. Since these loads seem to be popping Power Fets often, the thermal feedback is likely too slow to protect the FET, maybe installing a PTC resettable fuse may help save these devices.

Good luck with the mods.

Best,

[fubgumfaw]

I suspect the Zener is fine. What value is the gate series resistor? You could place a small cap on top as the lead cap, without a schematic and just guessing I would make the RC time constant ~100us. Also is there a series resistor to the CPU output? If so might consider placing a lead capacitor over this resistor as well. Since these loads seem to be popping Power Fets often, the thermal feedback is likely too slow to protect the FET, maybe installing a PTC resettable fuse may help save these devices.

Good luck with the mods.

Best,

As I was posting the above images, I was wondering what the value of that 1mm long resistor is too. Ummmm...does "tiny and black" count as a resistance value? haha! I'm betting it's 1K, but I have no idea its value. Based on 1K, 1nF ought to get close to 100uS.

A quick look on digikey and mouser didn't reveal any PTC's that are capable of around 120-200v and more than 7.5 amps. This is probably enough for higher voltages, but for 12-24v, it's not enough. 300 watts at 12v = 25 amps. I've already done this. I think a breaker like below is probably better since they are mostly voltage agnostic.

https://www.mouser.com/ProductDetail/TE-Connectivity-PB/W58-XB1A4A-30?qs=wvf%252BmaPDr9PS3kQIjLon9Q%3D%3D

This is the entire circuitry around the mosfet gate all the way to the CPU. I have not traced out the v-reg, but I think based on its traces that it connects to the load circuitry. I wonder if it is seeing too much variableness in its supply voltage to regulate well? I haven't traced out hardly anything here. It may be that those 2 small transistors are a gain circuit to boost a 3.3v logic level to 30ish volts? I haven't spent a lot of time trying to work out every detail here. I really wish they didn't grind off the tops of the IC's!

[mawyatt]

That resistor is probably 10~50 ohms I would think, just check with a DMM, in circuit measured value will be good. Those 5 pin two devices next to each other are probably not a transistor, but a simple amp or driver chip I would guess.

Best,

[Hamelec]

Hi, you may have a look to the circuit diagram posted here:
https://pic.mysku-st.ru/uploads/pictures/07/52/99/2020/11/23/e3773a.jpg
and maybe also read the complete thread
https://www.kirich.blog/obzory/izmeritelnoe/957-atorch-dl24-nedorogaya-elektronnaya-nagruzka-s-chetyrehprovodnym-podklyucheniem.html
Gate Resistor R2 is 1K, Gate driver is a LM321.

[fubgumfaw]

Hi, you may have a look to the circuit diagram posted here:
https://pic.mysku-st.ru/uploads/pictures/07/52/99/2020/11/23/e3773a.jpg
and maybe also read the complete thread
https://www.kirich.blog/obzory/izmeritelnoe/957-atorch-dl24-nedorogaya-elektronnaya-nagruzka-s-chetyrehprovodnym-podklyucheniem.html
Gate Resistor R2 is 1K, Gate driver is a LM321.

WOW!!! More info in a single haul than I ever expected! Thank you for posting this. Google translate to the rescue!

[mawyatt]

This is very good information. If that schematic is correct then there is a potential problem with the local feedback from the source which has a RC network with shunt cap to ground. This causes additional delay in the feedback path and with a 1K gate resistor and the gate capacitance (with Miller effect) likely the cause of instability under certain load conditions. Might be worth the effort to carefully analyze this circuit with the component values from the actual hardware. If we had this unit we could make some measurements and simulations to verify the stability.

Anyway this is where you might start looking for the root cause of the oscillations.

Best,

[beanflying]

Hi, you may have a look to the circuit diagram posted here:
https://pic.mysku-st.ru/uploads/pictures/07/52/99/2020/11/23/e3773a.jpg
and maybe also read the complete thread
https://www.kirich.blog/obzory/izmeritelnoe/957-atorch-dl24-nedorogaya-elektronnaya-nagruzka-s-chetyrehprovodnym-podklyucheniem.html
Gate Resistor R2 is 1K, Gate driver is a LM321.

Thanks for the links saves me some poking about today 

[frogblender]

This is very good information. If that schematic is correct then there is a potential problem with the local feedback from the source which has a RC network with shunt cap to ground. This causes additional delay in the feedback path and with a 1K gate resistor and the gate capacitance (with Miller effect) likely the cause of instability under certain load conditions. Might be worth the effort to carefully analyze this circuit with the component values from the actual hardware. If we had this unit we could make some measurements and simulations to verify the stability.

Anyway this is where you might start looking for the root cause of the oscillations.

+1

R2 = 1k seems a little high for driving a gate.   Big fat fets have big fat gate capacitance; the delay will be substantial.

[Hamelec]

This is very good information. If that schematic is correct then there is a potential problem with the local feedback from the source which has a RC network with shunt cap to ground. This causes additional delay in the feedback path and with a 1K gate resistor and the gate capacitance (with Miller effect) likely the cause of instability under certain load conditions. Might be worth the effort to carefully analyze this circuit with the component values from the actual hardware. If we had this unit we could make some measurements and simulations to verify the stability.

Anyway this is where you might start looking for the root cause of the oscillations.

Best,

I have done some simulations with Microcap 12, especially for the 1k gate R and found it ok.
But you are welcome also doing a simulation, because i am not so experienced..
R6/11=100R, R4/5=10k, R10=82k, R12=3k3, C4=39nF tbc
If anybody will figure out C1 and C5 would be great.

[beanflying]

I have some good LCR tweezers that will do in circuit so I will go over it tomorrow and get the values. Also had a new toy arrive today so I will drop the heatsink on mine and load it up a bit for a look at the Temperatures of the power bits in stock condition.

Really boring unloaded. But the back of the Fet was still well above ambient NL.

[Hamelec]

a nice toy! which Type? FLIR?

[beanflying]

Fluke TiS55 I am a newb  https://www.eevblog.com/forum/thermal-imaging/fluke-ti-series-cameras-and-software-specifically-tis55/msg3517756/#msg3517756

[beanflying]

Manuals and PC Software/Drivers. Fresh download as of 19/3/2021 to save the mediafire  one file at a time in the Zip file below.

Direct link to the Manuals and software area Link here to save the QR Code and further  if you want to check later for new versions https://www.mediafire.com/folder/1c04afq923397/A3

[mnementh]

snip

The shunts are 5w, R010 or 10 mOhms. I have tons of shunts because of upgrading BMS's. I've touched my shunts under load and they do warm up a little. Replacing them with bigger ones is probably a good idea. They are probably OK for 20 amps as is.

I have to admit...china makes some cheap stuff and often times that means cutting corners...and yet I keep going back for more china junk...LOL!

I did see what you had put down but even allowing for 5W being a typo your 20A 'ok' is flawed so I ran the numbers and looked and some basic spec sheets.

Damn you bean, you made me put set aside "The Sleevening" and drag out one of mine to look.

Resistors are 6332, so 1W nominal each, possibly higher if thick film metal. R010=0.01Ω, which is a normal value for the shunts we use in a 20A ESC. 2 in parallel is 0.005Ω. At 20A, that is 2W.

I think I may stick a heat sink on top of 'em just to be safe, but I'm gonna go for it when I have a chance rather than try and stack, recal, etc.

mnem
 

[beanflying]

Resistors are 0603's as I measured (6332 or 2512 are huge) them so more likely 1/8W or 'maybe' 1/4 but if you can save $0.0001 then Codeine ......

Stage one because I have the microscope out for my old eyes and the LCR Research Tweezers fired up in anger. Peripheral components and values as tested now onto the nitty gritty itty bitty ones.