Author Topic: Cheezeball DC Load: DL24P: Pump, or Dump ???  (Read 77806 times)

0 Members and 1 Guest are viewing this topic.

Offline frogblenderTopic starter

  • Regular Contributor
  • *
  • Posts: 124
Re: Cheezeball DC Load: DL24P: Pump, or Dump ???
« Reply #25 on: February 24, 2021, 02:28:59 pm »
Well...
I looked through my collection of shunts and I had none in 10 mOhms in anything other than 5 watts. I put in an order earlier at Mouser for 10w shunts and IXTH130N20T mosfets. I have half wave bridge diodes so that's not a problem. Copper heat sinks and plate are on the way. I'll be modding this thing soon enough.

This brings me to another application...and yes...of topic.
I don't really use mosfets in their linear region. I mess with motor controllers and BMS's more than anything else and they use mosfets where Rds matters. I need a current limiter. I've been using resistors for this, but mosfets are a better option. I build EV's for myself and for friends. I build my own chargers from PSU's. Charging a battery pack can happen at a much higher rate than the PSU's can deliver. Then they shut down to protect from over load. I've been making a bank of resistors that I can switch in or out that goes inline with the positive output wire from the PSU's. They get pretty warm, but this is OK for wire wound resistors.

The problem is 2 fold:
1. They are fixed resistance and not adjustable. I'd like to do more than just switch them in and out.
2. My charger is made of 8 Meanwell 24v PSU's in series. This lets me select an output voltage that is specific to various pack voltages. I need a resistor bank for each voltage output.

My thinking is a few large mosfets in parallel and running in their linear zone to act as variable resistors so I can limit the current with them like I do with the resistors.
Your ideal solution is a proper battery charging circuit, which will take care of a) initial trickle if needed, b) current-limited, but high-current bulk charge, and c) float charge at the end .... all of this tailored to whatever battery chemstry you are using.  If you are going to the trouble of monkeying fets to limit current, you should investigate proper battery chargers/charging circuits.

And I can guess that Elon's 250kW superchargers do NOT have resistors or other significantly lossy components in the charge path....  it would just waste too much power.
 
The following users thanked this post: mnementh

Offline mnementh

  • Super Contributor
  • ***
  • Posts: 17541
  • Country: us
  • *Hiding in the Dwagon-Cave*
Re: Cheezeball DC Load: DL24P: Pump, or Dump ???
« Reply #26 on: February 25, 2021, 07:43:32 pm »


Here you go. Cost: $37 plus they threw in my dry cleaning for free.  :-DD

mnem
alt-codes work here:  alt-0128 = €  alt-156 = £  alt-0216 = Ø  alt-225 = ß  alt-230 = µ  alt-234 = Ω  alt-236 = ∞  alt-248 = °
 

Offline mnementh

  • Super Contributor
  • ***
  • Posts: 17541
  • Country: us
  • *Hiding in the Dwagon-Cave*
Re: Cheezeball DC Load: DL24P: Pump, or Dump ???
« Reply #27 on: March 02, 2021, 10:42:47 pm »
We talked about MOSFETs for DC loads the other day. I happened across the FDL100N50 today while looking for something else.

With a good enough heatsink (liquid cooled maybe) it has a DC SOA good for at least 1500W within 100A and 500V. Rjc is 0.05 deg C/W  so a 25 deg C heatsink actually has some headroom at 1500W. It's pretty linear with about 20mV/a gate sensitivity.

https://www.onsemi.com/pdf/datasheet/fdl100n50f-d.pdf

Robert763 passed this on in our TEAnonymous thread. At first I thought this part was like bringing a tank to a knife fight... but after looking at the much broader SOA right around where I want to operate combined with the much higher VGSS, I'm feeling like it might not be a bad idea to invest in this $25 (delivered) part to upgrade a $30 toy electronic load.

The primary failure mode (aside from ding-dongs just not paying attention to the PD) we've seen caused by excessive VGSS from a weakness in the original design makes me feel a lot better aboot paying an extra tenner for that extra overhead.

   

Next question becomes how to suck that heat away... I'm seriously considering investing in a new AiO Water cooler for my gaming PC and kicking the original one down to this project.  :o   Would be a lot better than this CoolerMaster I had lying around, I suspect.

mnem
 :-/O
« Last Edit: March 03, 2021, 03:16:05 pm by mnementh »
alt-codes work here:  alt-0128 = €  alt-156 = £  alt-0216 = Ø  alt-225 = ß  alt-230 = µ  alt-234 = Ω  alt-236 = ∞  alt-248 = °
 

Offline frogblenderTopic starter

  • Regular Contributor
  • *
  • Posts: 124
Re: Cheezeball DC Load: DL24P: Pump, or Dump ???
« Reply #28 on: March 02, 2021, 11:36:45 pm »
I happened across the FDL100N50 today while looking for something else.
https://www.onsemi.com/pdf/datasheet/fdl100n50f-d.pdf

Kudos to Robert763 for digging up the FDL100N50 .... this has 4x (4x!) better Rjc (.05°C/W, vs .2°C/W) than the part I used (and I looked for good Rjc.... must've done a ManScan).   That's a bit of a game-changer.


 
The following users thanked this post: mnementh

Offline fubgumfaw

  • Regular Contributor
  • *
  • Posts: 104
  • Country: us
Re: Cheezeball DC Load: DL24P: Pump, or Dump ???
« Reply #29 on: March 08, 2021, 07:39:48 am »
I have the heat sink beefed up considerably and turned it up to 500 watts.
I'm still waiting on the zener diodes, better mosfets and larger diode to arrive.
I've tried it on the factory mosfet, which blew instantly on 48v and 10 amps.
I then dropped in a 100v mosfet I had and tried at 1 amp and it too blew immediately.
I'm pretty sure the 15v zener will reign in the gate voltages that are killing mosfets.

1. 1/4" thick copper plate on the mosfet and diode.
2. 4 solid copper CPU heat sinks soldered to the copper plate.
3. Factory fan mounted on top of all of that.
4. Probably 5X more heat sinking than the factory heat sink.
5. The heatsink is now 100% electrically isolated from everything else.
6. Support for the heavy heat sink is at the corners and through the screw hole in the mosfet.
7. The board is no longer bowed around the mosfet.



« Last Edit: March 17, 2021, 10:11:37 am by fubgumfaw »
 

Offline fubgumfaw

  • Regular Contributor
  • *
  • Posts: 104
  • Country: us
Re: Cheezeball DC Load: DL24P: Pump, or Dump ???
« Reply #30 on: March 08, 2021, 04:50:33 pm »
I replaced the blown factory mosfet with an HY5110 and it blew almost immediately. I also burnt the trace on the source leg back to the screw terminal and the 2 small shunts. Bigger shunts will be here soon. Mouser has delayed my order twice now. The trace is repaired and beefed up with some solder braid. I'll pull the screw terminal out and add some copper sheet on the current path trace back to that. This ought to hold up now. I'm highly skeptical that the DIY version could handle 1000 watts. I know my mosfet died shorted which created a massive current path across my test LIPO, but still...the traces with no reinforcing and that thin copper and those tiny shunts...not going to happen!
 

Offline frogblenderTopic starter

  • Regular Contributor
  • *
  • Posts: 124
Re: Cheezeball DC Load: DL24P: Pump, or Dump ???
« Reply #31 on: March 08, 2021, 05:00:08 pm »
I replaced the blown factory mosfet with an HY5110 and it blew almost immediately. I also burnt the trace on the source leg back to the screw terminal and the 2 small shunts. Bigger shunts will be here soon. Mouser has delayed my order twice now. The trace is repaired and beefed up with some solder braid. I'll pull the screw terminal out and add some copper sheet on the current path trace back to that. This ought to hold up now. I'm highly skeptical that the DIY version could handle 1000 watts. I know my mosfet died shorted which created a massive current path across my test LIPO, but still...the traces with no reinforcing and that thin copper and those tiny shunts...not going to happen!

Do you have the zener in there?  If no, you will continue to blow fets
 

Offline fubgumfaw

  • Regular Contributor
  • *
  • Posts: 104
  • Country: us
Re: Cheezeball DC Load: DL24P: Pump, or Dump ???
« Reply #32 on: March 08, 2021, 05:13:38 pm »
I replaced the blown factory mosfet with an HY5110 and it blew almost immediately. I also burnt the trace on the source leg back to the screw terminal and the 2 small shunts. Bigger shunts will be here soon. Mouser has delayed my order twice now. The trace is repaired and beefed up with some solder braid. I'll pull the screw terminal out and add some copper sheet on the current path trace back to that. This ought to hold up now. I'm highly skeptical that the DIY version could handle 1000 watts. I know my mosfet died shorted which created a massive current path across my test LIPO, but still...the traces with no reinforcing and that thin copper and those tiny shunts...not going to happen!

Do you have the zener in there?  If no, you will continue to blow fets

Mouser keeps delaying my order. I have IXTH130N20T mosfets, 10w shunts, and 1.3w 15v zeners coming.
I know about the gate voltage issues...hence the zeners.
I'm betting blowing the source leg trace and puny shunts is partly from the blown mosfets from excessive gate voltages. OF course, I did go straight to 500 watts and that was not the best way to test things either! I should have tried it out at 185 watts first. :)
 

Offline fubgumfaw

  • Regular Contributor
  • *
  • Posts: 104
  • Country: us
Re: Cheezeball DC Load: DL24P: Pump, or Dump ???
« Reply #33 on: March 09, 2021, 03:13:59 am »
Well...
I looked through my collection of shunts and I had none in 10 mOhms in anything other than 5 watts. I put in an order earlier at Mouser for 10w shunts and IXTH130N20T mosfets. I have half wave bridge diodes so that's not a problem. Copper heat sinks and plate are on the way. I'll be modding this thing soon enough.

This brings me to another application...and yes...of topic.
I don't really use mosfets in their linear region. I mess with motor controllers and BMS's more than anything else and they use mosfets where Rds matters. I need a current limiter. I've been using resistors for this, but mosfets are a better option. I build EV's for myself and for friends. I build my own chargers from PSU's. Charging a battery pack can happen at a much higher rate than the PSU's can deliver. Then they shut down to protect from over load. I've been making a bank of resistors that I can switch in or out that goes inline with the positive output wire from the PSU's. They get pretty warm, but this is OK for wire wound resistors.

The problem is 2 fold:
1. They are fixed resistance and not adjustable. I'd like to do more than just switch them in and out.
2. My charger is made of 8 Meanwell 24v PSU's in series. This lets me select an output voltage that is specific to various pack voltages. I need a resistor bank for each voltage output.

My thinking is a few large mosfets in parallel and running in their linear zone to act as variable resistors so I can limit the current with them like I do with the resistors.
Your ideal solution is a proper battery charging circuit, which will take care of a) initial trickle if needed, b) current-limited, but high-current bulk charge, and c) float charge at the end .... all of this tailored to whatever battery chemstry you are using.  If you are going to the trouble of monkeying fets to limit current, you should investigate proper battery chargers/charging circuits.

And I can guess that Elon's 250kW superchargers do NOT have resistors or other significantly lossy components in the charge path....  it would just waste too much power.


I agree about the heat/waste and this is a good reason to move away from the "resistors as current limiters" solution I've used previously. Google isn't helping a lot and I've search eevblog to no avail too. Really all I care about is limiting current at 25 amps and a minimum of 36v and a maximum of 150v. All the other typical charging stuff is already taken care of. I don't care about chemistry specific charging. That's handled too. Just need 36v to 150v at 25 amps max. I'm looking for something that would sense current and adjust the gate voltage so the mosfets act as high current variable resistances. This will create heat since the mosfets will be in their linear region. There are other limitations to the design that need to keep the supply voltages set where I put them and just throttles current.

This is a thread I started on it:
https://www.eevblog.com/forum/projects/power-supply-current-limiting-issue/
 

Offline mnementh

  • Super Contributor
  • ***
  • Posts: 17541
  • Country: us
  • *Hiding in the Dwagon-Cave*
Re: Cheezeball DC Load: DL24P: Pump, or Dump ???
« Reply #34 on: March 11, 2021, 10:45:58 pm »
I replaced the blown factory mosfet with an HY5110 and it blew almost immediately. I also burnt the trace on the source leg back to the screw terminal and the 2 small shunts. Bigger shunts will be here soon. Mouser has delayed my order twice now. The trace is repaired and beefed up with some solder braid. I'll pull the screw terminal out and add some copper sheet on the current path trace back to that. This ought to hold up now. I'm highly skeptical that the DIY version could handle 1000 watts. I know my mosfet died shorted which created a massive current path across my test LIPO, but still...the traces with no reinforcing and that thin copper and those tiny shunts...not going to happen!

Yeeah... Did you see the $200 MOSFET module they recommended for the "DIY 1500W" version...? Obvi not gonna bolt that directly to the PCB; would have to build it offboard with shunts offboard as well as all the power-handling DC busses. This unit would only be used to drive that module.  ::)

mnem
*still waiting on mine*
alt-codes work here:  alt-0128 = €  alt-156 = £  alt-0216 = Ø  alt-225 = ß  alt-230 = µ  alt-234 = Ω  alt-236 = ∞  alt-248 = °
 

Offline fubgumfaw

  • Regular Contributor
  • *
  • Posts: 104
  • Country: us
Re: Cheezeball DC Load: DL24P: Pump, or Dump ???
« Reply #35 on: March 12, 2021, 08:50:56 pm »
Did you see the $200 MOSFET module they recommended for the "DIY 1500W" version...? Obvi not gonna bolt that directly to the PCB; would have to build it offboard with shunts offboard as well as all the power-handling DC busses. This unit would only be used to drive that module.  ::)

I didn't see this! Can you find it please and post the URL? I'm all for seeing what they did and maybe even copying it. I looked for it on aliexpress...nada.

I got the zener and shunts and put them on the board with much reinforced power traces. I know I have the 15v zener in place correctly across the gate and source. I connected to a fully charged 16Ah 3S LIPO pack (12.6v) and started at 1 amp...all good. I pushed the amps up higher and got to 10 amps...crackle fizzle! I pulled the mosfet and it was fried. I have several HY5110 mosfets so I don't care if I kill a few. I put in another one and the gate drive is cooked too! Now I might need to buy another one of these things! I haven't spent any time looking at the small components yet that drive the mosfet. Hopefully its just the output component that is cooked. Its a 5 legged part...probably a little mosfet is my guess. The zener is OK. I got a 1.5w version just in case.

I ran .3mm copper sheet to the pins on the connector and to all the traces that carry current. Even the bare area under the mosfet got added copper since this goes to the drain leg solder pad. I'm pretty sure the traces can handle 50 amps now! It's all soldered down after scraping off the solder resist as needed.

I'm thinking a diode between the driver and the mosfet gate would have protected it from the main mosfet killing whatever died. However, that would also limit the gate voltage to the break over voltage of the diode. That may not be an issue since the gate voltage can swing so high!

After I get it working again and it's reliable running a single mosfet, I'll add a second mosfet to the bottom of the board. I need to make an aluminum plate for the whole thing to mount onto and that will make a nice heat sink for the second mosfet.
« Last Edit: March 12, 2021, 09:29:25 pm by fubgumfaw »
 

Offline fubgumfaw

  • Regular Contributor
  • *
  • Posts: 104
  • Country: us
Re: Cheezeball DC Load: DL24P: Pump, or Dump ???
« Reply #36 on: March 12, 2021, 09:28:34 pm »
The ATORCH girl is certainly easy on the eyes

https://ibb.co/QCL0JRr

I bought 2 of the DIY version on aliexpress...so cheap and why not? I've done all the hard part already.
https://www.aliexpress.com/item/1005001565062996.html?spm=a2g0s.9042311.0.0.267f4c4d4YCMIK
« Last Edit: March 12, 2021, 10:20:36 pm by fubgumfaw »
 

Offline mnementh

  • Super Contributor
  • ***
  • Posts: 17541
  • Country: us
  • *Hiding in the Dwagon-Cave*
Re: Cheezeball DC Load: DL24P: Pump, or Dump ???
« Reply #37 on: March 15, 2021, 04:00:57 pm »
Did you see the $200 MOSFET module they recommended for the "DIY 1500W" version...? Obvi not gonna bolt that directly to the PCB; would have to build it offboard with shunts offboard as well as all the power-handling DC busses. This unit would only be used to drive that module.  ::)

I didn't see this! Can you find it please and post the URL? I'm all for seeing what they did and maybe even copying it. I looked for it on aliexpress...nada....



It's right there in the Q&A, if you read it... ;) They say to get up into the 1KW range, you'll need to upgrade the pass element to something more powerful, like a VMO400-02F MegaMOSFET module. ONLY ~$200 list price @ DigiKey.  :palm:

That's as much CYA as actual enginerding, I'll wager. By recommending something with this arbitrarily weird packaging, it places ALL the onus upon the user to ensure correct gate drive, current-limiting, shunting for the measurement circuit, crowbar zeners where appropriate etc... They can claim "up to 999 Watts!!!" without actually having to back it up in any way.  |O

In Related News... My two cheap & cheerful DC load kits have arrived; with multi-package battery trays and USB adaptor PCBs amazingly intact as they were loose inside the padded baggie.

They both have IRFP260N pass transistors; these appear to be genuine IR brand, but reclaimed parts. So the claims of "shorted transistor new out the box" I've read are not necessarily bogus; these tend to be the most common point of failure in any kind of equipment that uses them for obvious reasons. I recommend removing and testing the FET with your Cheap Chinese Component Tester and doing the recommended crowbar zener and maybe adding a fuse mod before actually applying power to this device.

I intend to do this, as well as paralleling the two OEM IRFP260Ns that came with on one single unit, when I finish my current "PC rebuild in a new case" project.

mnem
 :bullshit:
« Last Edit: March 15, 2021, 05:35:38 pm by mnementh »
alt-codes work here:  alt-0128 = €  alt-156 = £  alt-0216 = Ø  alt-225 = ß  alt-230 = µ  alt-234 = Ω  alt-236 = ∞  alt-248 = °
 

Offline fubgumfaw

  • Regular Contributor
  • *
  • Posts: 104
  • Country: us
Re: Cheezeball DC Load: DL24P: Pump, or Dump ???
« Reply #38 on: March 15, 2021, 06:21:03 pm »


They both have IRFP260N pass transistors; these appear to be genuine IR brand, but reclaimed parts. So the claims of "shorted transistor new out the box" I've read are not necessarily bogus; these tend to be the most common point of failure in any kind of equipment that uses them for obvious reasons. I recommend removing and testing the FET with your Cheap Chinese Component Tester and doing the recommended crowbar zener and maybe adding a fuse mod before actually applying power to this device.

mnem
 :bullshit:

The  IRFP260N is not a great choice for this application. The package is only rated for 300 watts and 175C. It will get pretty warm.

It must be specific sales threads on aliexpress that list the IXYS mosfet. I bought from here and I don't see the mosfet listed in the comments or elsewhere. Oh well...don't really care I guess. I'll just use 2 TO-247 mosfets in parallel. That ought to get me plenty of current capability.
https://www.aliexpress.com/item/1005001565062996.html?spm=a2g0s.12269583.0.0.24527e8eBUrMDJ

I googled for the part number on it. This is what I found:
https://pdf1.alldatasheet.com/datasheet-pdf/view/126167/IXYS/VMO400-02F.html

However going straight to the IXYS/Littlefuse web site did not find this device. Mouser didn't have it either. Maybe its a fake part or some Chinesium knock-off?
Regardless, it's just a big mosfet in there...

I noticed this too. The original mosfet was not what I would call "new looking". I did test it on a component tester and it seemed OK so I used it as is. The first capacity tester I got, the mosfet did NOT test OK. I sent it back! Those inexpensive Chinese made component testers are super handy! I've vetting thousands of used parts with them.

A fuse...that's a good idea! After you beef up the traces, better heat sinking, zener, larger shunts, etc...a fuse would be a good add-on too.
Before I killed the CPU output for the mosfet, I had just gone to 10 amps on a 3S LIPO pack or about 125 watts. The LCD never showed more than 10 amps so I was under the 180 watt limit. Maybe even with the zener in place, amp draw peaked way over this? Before I beefed up the traces, the drain trace had already melted through right at the leg. I also cooked the 2 tiny shunts. A fuse or reset-able breaker may have saved it from further damage after I removed the other weak spots.

I have 2 of the DIY boards coming. I tried to repair mine, but the driver circuitry is blown back inside the CPU. I think because semi's tend to blow to dead shorts that a way to protect the gate driver might be worth while. This is highly likely what killed mine. I think even a silicon diode would do the job. Put it between the mosfet gate and the driver circuit. This will limit the lower voltage some to over .7v, but I think that will be OK. The IXTH130N20T needs 5v to get to Vgsth so losing a little voltage won't hurt much by adding a diode inline with the gate.
« Last Edit: March 15, 2021, 07:42:07 pm by fubgumfaw »
 
The following users thanked this post: mnementh

Offline mnementh

  • Super Contributor
  • ***
  • Posts: 17541
  • Country: us
  • *Hiding in the Dwagon-Cave*
Re: Cheezeball DC Load: DL24P: Pump, or Dump ???
« Reply #39 on: March 15, 2021, 08:01:39 pm »


It's right here at the bottom of the page you linked to.  :-// The VMO400-02F is listed at DigiKey, at a cost of ~$200 delivered in single Qty. That is where I got the datasheet I attached to the bottom of my post.

Really, dude... read all the way to the bottom before you freak out.  :-DD

I'm not looking for high voltage. The SOA for the IRFP260N sez it is good for 20A at the voltage I want to use it for. I'm planning to parallel 2 in my application, and this is just to see if it will work before I buy a $25 part.

Empirical engineering does not mean not taking precautions. ;)

mnem
 :-/O

alt-codes work here:  alt-0128 = €  alt-156 = £  alt-0216 = Ø  alt-225 = ß  alt-230 = µ  alt-234 = Ω  alt-236 = ∞  alt-248 = °
 

Online beanflying

  • Super Contributor
  • ***
  • Posts: 7355
  • Country: au
  • Toys so very many Toys.
Re: Cheezeball DC Load: DL24P: Pump, or Dump ???
« Reply #40 on: March 16, 2021, 04:32:09 am »
Pile of bits arrived today but time to play not yet  :palm: Love the junk power supply and the adapter it came with NEITHER suits our Aussie Angled pins :horse: The Battery and USB adapters look ok but no way I would suggest they be pushed anywhere near 150W more like a few Amps maximum looking at the traces.

Plan is to drop the Cooler of my existing 150W load and put something out of my PC Salvage box on it for a 'spare' lower power load at this stage. Also now I have the actual new board here I should be able to fit in some CAD time for an enclosure design to keep some of the dust and fingers out.

EDIT

Couple of things maybe of note the old one ran easily off a 5V USB wall supply (Micro USB jack at the rear of the board) I have installed in the shack and so does the new one. Both were labelled on the PCB as 6-12V supply but initial inspection shows the new one fine at 5V.

Minor niggle the old one had two fan jacks the new one only has a single must have saved them $0.03 per item  :palm:

Both boards share common sizes and Fan and feet mounting point locations.

5pm so it is Beer O'clock  :popcorn:

Edit of the Edit over a Beer  ;D

Enclosure rough out. Not to dimensioned accurately or corners rounded etc but 8 bits of 3mm Acrylic is the plan.
« Last Edit: March 16, 2021, 07:17:59 am by beanflying »
Coffee, Food, R/C and electronics nerd in no particular order. Also CNC wannabe, 3D printer and Laser Cutter Junkie and just don't mention my TEA addiction....
 

Online tautech

  • Super Contributor
  • ***
  • Posts: 28058
  • Country: nz
  • Taupaki Technologies Ltd. Siglent Distributor NZ.
    • Taupaki Technologies Ltd.
Re: Cheezeball DC Load: DL24P: Pump, or Dump ???
« Reply #41 on: March 16, 2021, 07:38:14 am »
Edit of the Edit over a Beer  ;D

Enclosure rough out. Not to dimensioned accurately or corners rounded etc but 8 bits of 3mm Acrylic is the plan.
Acrylic is flammable isn't it ?  :P


Burn baby burn.......  :-DD
Avid Rabid Hobbyist
Siglent Youtube channel: https://www.youtube.com/@SiglentVideo/videos
 
The following users thanked this post: mnementh

Online beanflying

  • Super Contributor
  • ***
  • Posts: 7355
  • Country: au
  • Toys so very many Toys.
Re: Cheezeball DC Load: DL24P: Pump, or Dump ???
« Reply #42 on: March 16, 2021, 07:40:08 am »
Edit of the Edit over a Beer  ;D

Enclosure rough out. Not to dimensioned accurately or corners rounded etc but 8 bits of 3mm Acrylic is the plan.
Acrylic is flammable isn't it ?  :P


Burn baby burn.......  :-DD

If the enclosure is burning you already have a fire problem ;)
Coffee, Food, R/C and electronics nerd in no particular order. Also CNC wannabe, 3D printer and Laser Cutter Junkie and just don't mention my TEA addiction....
 
The following users thanked this post: mnementh

Online Brumby

  • Supporter
  • ****
  • Posts: 12281
  • Country: au
Re: Cheezeball DC Load: DL24P: Pump, or Dump ???
« Reply #43 on: March 16, 2021, 08:26:14 am »
Just looking at the "shunts" on my board.   :palm:

Some serious beefing up work required for any serious application.

So, lesseee...
 * Beef up shunts
 * Beef up tracks
 * Zener
 * MOSFET upgrade

.... and I still have to dig out that old CPU cooler.

Have I missed anything....?
 
The following users thanked this post: mnementh

Online tautech

  • Super Contributor
  • ***
  • Posts: 28058
  • Country: nz
  • Taupaki Technologies Ltd. Siglent Distributor NZ.
    • Taupaki Technologies Ltd.
Re: Cheezeball DC Load: DL24P: Pump, or Dump ???
« Reply #44 on: March 16, 2021, 08:38:51 am »
Fire extinguisher ?
Avid Rabid Hobbyist
Siglent Youtube channel: https://www.youtube.com/@SiglentVideo/videos
 
The following users thanked this post: mnementh

Online beanflying

  • Super Contributor
  • ***
  • Posts: 7355
  • Country: au
  • Toys so very many Toys.
Re: Cheezeball DC Load: DL24P: Pump, or Dump ???
« Reply #45 on: March 16, 2021, 09:45:33 am »
Just looking at the "shunts" on my board.   :palm:

Some serious beefing up work required for any serious application.

So, lesseee...
 * Beef up shunts
 * Beef up tracks
 * Zener
 * MOSFET upgrade

.... and I still have to dig out that old CPU cooler.

Have I missed anything....?

For a start it makes sense to undo the FET and power side of the board and look at some numbers and characterize what a 'sensible' set of numbers are for it. Simply looking at the Maximum/Peak  :bullshit: of the component without looking at the rest is a bit pointless like complaining it wont do A,B or C ambit Voltage/Current/Power or it failed because X was inside the figure for the FET and ignoring what else is around it.

My rough look at the old board and track weight and width and I set a 'guesstimate' of 10A being a sensible figure for  the PCB to handle so 15V at about 10A for its 150W. Trying to push 30A@5V would certainly see a meltdown even though the old shunts might have survived. The new shunts if current split might actually be ok at 5A each without knowing how they fit in.

Gut feel says if you want more than 10A (maybe a bit more when the numbers are done) out of this new one you need to be making an entire new Power end and taking the Fet and Cooler to a standalone system.

Tomorrow Job I hope if time permits.
Coffee, Food, R/C and electronics nerd in no particular order. Also CNC wannabe, 3D printer and Laser Cutter Junkie and just don't mention my TEA addiction....
 
The following users thanked this post: mnementh

Offline mnementh

  • Super Contributor
  • ***
  • Posts: 17541
  • Country: us
  • *Hiding in the Dwagon-Cave*
Re: Cheezeball DC Load: DL24P: Pump, or Dump ???
« Reply #46 on: March 16, 2021, 01:34:59 pm »
My gut is scaly.  :-DD

But it generally agrees with your estimate of the original IRFP260N FET. Current capacity of the PCB... not so much. I've seen much less copper fill on 30-40A ESCs, so the main fear I have in my application is what is the actual wattage capacity of those cheap SMD shunts vs what they're rated for. OTOH, as I'm already going to be doing mod work for the 2nd pass transistor, what's the harm in laying down some extra copper strands over the existing traces...?  :-//

mnem
 :bullshit:
« Last Edit: March 16, 2021, 06:50:50 pm by mnementh »
alt-codes work here:  alt-0128 = €  alt-156 = £  alt-0216 = Ø  alt-225 = ß  alt-230 = µ  alt-234 = Ω  alt-236 = ∞  alt-248 = °
 

Online beanflying

  • Super Contributor
  • ***
  • Posts: 7355
  • Country: au
  • Toys so very many Toys.
Re: Cheezeball DC Load: DL24P: Pump, or Dump ???
« Reply #47 on: March 17, 2021, 04:16:42 am »
Ok Shunts because they are easy 2 parallel 0603 1/8W (most likely) 0.01 \$\Omega\$ shunts. So 1/4W peak for the pair. Take my 10A guesstimate for the tracks and you will likely blow the snot out of the SMD shunts as the pair would handle circa 7A at maximum current :palm:

So yep upgrade needed for a start :-+ Quick check of 0805 0.01's in parallel would get you to 14A at a combined 1/2W or larger if you have or can get them easily.



« Last Edit: March 17, 2021, 07:47:02 am by beanflying »
Coffee, Food, R/C and electronics nerd in no particular order. Also CNC wannabe, 3D printer and Laser Cutter Junkie and just don't mention my TEA addiction....
 

Offline fubgumfaw

  • Regular Contributor
  • *
  • Posts: 104
  • Country: us
Re: Cheezeball DC Load: DL24P: Pump, or Dump ???
« Reply #48 on: March 17, 2021, 04:29:21 am »
Ok Shunts because they are easy 2 parallel 0603 1/8W (most likely) 0.01 \$\Omega\$ shunts. So 1/4W peak for the pair. Take my 10A guesstimate for the tracks and you will likely blow the snot out of the SMD shunts as the pair would handle circa 7A at maximum current :palm:

So yep upgrade needed for a start :-+

I posted all of this already. Shunts will burn up. Same for the traces to drain and source legs. I added copper to all of this and used 10w .01ohm shunts. My copper goes direct to the pins on the 4 terminal connector and all the current paths. Add a fuse or small breaker inline so you have a safety device in there too.
 

Online beanflying

  • Super Contributor
  • ***
  • Posts: 7355
  • Country: au
  • Toys so very many Toys.
Re: Cheezeball DC Load: DL24P: Pump, or Dump ???
« Reply #49 on: March 17, 2021, 04:37:59 am »

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.
Coffee, Food, R/C and electronics nerd in no particular order. Also CNC wannabe, 3D printer and Laser Cutter Junkie and just don't mention my TEA addiction....
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf