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

[Pukker]

Can anyone tell me the specs of the external temp. sensor.
Which value NTC en wich curve.
I can measure the R value of course.
Is it: Ntc 5K 10K 15K 20K 50K 100K 1M
and is it curve 3950 or 3435 or 3470?

Thank you in advance.

[stts]

Can anyone tell me the specs of the external temp. sensor.
Which value NTC en wich curve.
I can measure the R value of course.
Is it: Ntc 5K 10K 15K 20K 50K 100K 1M
and is it curve 3950 or 3435 or 3470?

Thank you in advance.

Its an equation you pump measured values into. It comes out to 3600 beta for the long wire probe that measures 10k ohm at 25C. I put 10 tiny SMD chips of those in my basket for 35cents each to order on my next parts aquisition. I have micro wire and that amber tape to make up more of those thermo sensors.

Also, I ran the unit for 260 hours with no troubles at the low current that my supply puts out. I realized later that the unit only times to 200 hours before it resets to zero and counts over. But that should cover most any bodies testing demands.

In addition, I ordered samples of the pricey IXTK90n25L2 from Littlefuse direct. I requested 4 of them and can only hope they turn out to be free samples. They go to my work address at the plant so I am expecting exciting things. But may end up with a can of worms, hah. It just depends on how nosey the sales people are. I saw they have these much cheaper on Ebay, but the pictures look like they are all fakes. Another company in Honk Kong has them rather cheap as well at not too hatefull $10. But I dont know what shipping would be. It seems like a more reputable parts company, which would make sense from a Hong Kong establishment. I bet direct from china will mostly be fakes.

In the mean time, I will use my tester to test the faulty steering driver for a fork lift. Nothing is shorted but drive quits under moderate load. This thing will sink plenty of load  that I can increase incrimentally to test all of the mosfets in the drive. This cant test them all at once but it can easily test them individually to see which one is weak. And the best part is its in circuit testing. Saves me a ton of work. It would take so much heat to remove all those mosfets that it would be better to replace them all because of the heat stress uncertainty.

[Pukker]

@STTS, Sorry, can't read an answer to my qeustion in your reply.

[stts]

@STTS, Sorry, can't read an answer to my qeustion in your reply.

"It comes out to 3600 beta for the long wire probe that measures 10k ohm at 25C."

Sorry for saying all that other stuff that left you confused.   

[Pukker]

@STTS, Sorry, can't read an answer to my qeustion in your reply.

"It comes out to 3600 beta for the long wire probe that measures 10k ohm at 25C."

Sorry for saying all that other stuff that left you confused.   

Found it. NTC is 10k Beta 3435
This sensor is standard to buy in many configurations at AliExpress etc.

[Pukker]

FYI.
Measuring the temperature of the STPS41H100CG diode, It gives 75 degrees Celsius at 6 Amp. (12 Volt)
Datasheet reads an Tj of 175 degrees, but I don't like that kind of heating for diode and also not for the PCB.
Atorch specs for the load are 20 Amp. max, but I don't know how they would do that.
I think I put some heatsink on the diode.

( BTW: The temperature of the Mosfet reads 35 degrees Celsius, I have an good cooler.)

The load comes from China, so you never know the STPS41H100CG is original.
I bought an new one without fan. I replaced the Mosfet with an original one.
I observed the Mosfet I removed and that one had some serious markings that shows
it was mounted with an screw before. So I think it was an used one.

[hpapagaj]

Anyone knows what is the type of display which is used on DL24? I accidentally broke it. 

[radeko]

I have problem with DL24P.
 I placed a zener diode between the gate and source of the mosfet, but it did not protect the mosfet but on the contrary it damaged the mosfet and the opamps.
Could someone show a picture of how they placed the zener diode? Black side towards the gate?
I connected the black side of the zener diode to the gate and the other side to the source.

[Pukker]

I have problem with DL24P.
 I placed a zener diode between the gate and source of the mosfet, but it did not protect the mosfet but on the contrary it damaged the mosfet and the opamps.
Could someone show a picture of how they placed the zener diode? Black side towards the gate?
I connected the black side of the zener diode to the gate and the other side to the source.

Here an picture (by an another member of this forum)
how to place the zener.
As you described it, you have done it OK.

[radeko]

It is another solution. Double zener.

[Pukker]

It is another solution. Double zener.

Don't know if it is posible there can be negative voltages on the gate,
but your option will also work I think.

[Tangent_Tracker]

Folks I thought I'd have a heatsink that fits this but it seems mine are either too old or not compatible. Anyone any cheap suggestions?

[Pukker]

Folks I thought I'd have a heatsink that fits this but it seems mine are either too old or not compatible. Anyone any cheap suggestions?

Search for Socket 775 compatible processor coolers.
Internet, computershops, second-hand market.
That socket is a bit old and much used, so should be to find.
Also for acceptable cost to find on Ali and Ebay.

[Filippo52]

Ho usato un dissipatore di calore Intel LGA1151 che ha la distanza tra i fori esattamente la stessa della scheda pcb DL24, ovvero 75 mm.
LGA775 ha una distanza tra i fori di 72 mm e richiede quindi un minimo di adattamento da parte tua.
Entrambi i dissipatori sono uguali e non creano problemi con le prese sulla scheda DL24; tuttavia la ventola LGA1151 gira di più quindi, anche se fa un po' più di rumore, dissipa meglio il calore.
In entrambi i casi, però, hai il problema che qui si è visto più volte.


Usando gli attacchi originali il dissipatore rimane a 3-4 mm dalla superficie del mosfet e quindi non funziona. Devi togliere gli attacchi originali e pensare ad usare viti e dadi, io ho comprato un supporto ventola di raffreddamento per fare una cosa migliore.
Comunque per avere un miglior contatto tra la base del dissipatore e il mosfet visto che si ha a disposizione un bel foro passante, ho praticato un foro alla base del dissipatore (Photo 3) e creato in questo modo la sede per una vite autofilettante l'aderenza del mosfet è garantita contro eventuali torsioni della scheda.
Allego alcune figure





Nella foto 4 potete vedere due piccoli dissipatori che ho messo sui diodi di protezione schottky e sulle resistenze di shunt.


Il risultato è buono con 12 volt e 9 A il mosfet rimane sotto i 50°C.
L'ho aumentato fino a 110w in modalità DC e non sono riuscito a vedere alcuna forma di oscillazione al gate del mosfet con il mio oscilloscopio: solo una tensione DC pulita. Al momento non ho quindi messo zener.
Qualcuno può dirmi se l'oscillazione appare in altri modi? ad esempio CV...CR ?
Grazie

[SpottedDick]

A new model has entered the scene!

The fact that this now has a fuse has me hopeful that they've gone and fixed all the original flaws (ringing at the gate etc).

However, (a big however), this feels way overpriced. The original board was/is €22. This one is *€65*!!!

Anyone brave enough to buy it and scope it, see if it's any better?

https://www.aliexpress.com/item/1005004773128494.html

EDIT: Does anyone know the purpose of the large green resistors on the new board?

[BILLPOD]

Ho usato un dissipatore di calore Intel LGA1151 che ha la distanza tra i fori esattamente la stessa della scheda pcb DL24, ovvero 75 mm.
LGA775 ha una distanza tra i fori di 72 mm e richiede quindi un minimo di adattamento da parte tua.
Entrambi i dissipatori sono uguali e non creano problemi con le prese sulla scheda DL24; tuttavia la ventola LGA1151 gira di più quindi, anche
I'ma no capeesh

[Filippo52]

there is also this quite recent

https://tinyurl.com/3kzpaa4r

which has the same type of Hw with large green resistors and above all uses 4 power mosfets, even if in my opinion the heat sink is too small, and it is a vertical execution.
Both have the Hw to connect to other similar specimens and make one large electronic load, at this point very bulky and very expensive
However, this too costs a lot and is not suitable for those who want to DIY with customized versions.

[Filippo52]

This is a new execution of the DL24 Pcb with a cooler, also Intel, but larger: It was once intended for Bloomfield CPUs with socket 1366.
As you can see from the photo,



the dimensions are much more generous (it comes out from the edges of the card) and the weight is much higher.
In this case it is necessary to make some modifications to the card:
1 - make additional holes because socket 1366 had a distance between the holes of 80mm and not 75.
2 - intervene on the dissipator supports because they hit the board power supply sockets and the output sockets for measuring the current.
3 - remove part of the support from one leg of the heatsink to prevent it from overlapping the blue bluetooth leds
4 - change the power socket of the Intel heatsink fan because with this Heatseink there is little space between the heatsink and the power socket on board the card

The dissipation power was suitable for CPUs with a TDP of even 150W and therefore similar to what we need.

I did this because the previous used cooler of the INTEL socket 115X series was suitable for 60-90W CPUs and therefore not really suitable for our purposes if we want to push the limits of this board without touching the components.

I would like to point out that in the last two cards that I have used, the Mosfet that arrives when installed is the IRFP264 that I have disassembled and tested separately. The weight is 5.6 g and the RDS is 0.037 ohm; and these are values which suggest that it is an authentic one and not a counterfeit mosfet.

So far I have also tested this up to 120W and it stays at 42°C with the NTC of the board,
However touching the back of the card the part near the solder of the "DRAIN" gets very hot and it's something I don't like.
I fear that there is some heat in the mosfet that is not dissipated by the Mosfet Leadframe but passes through the DRAIN Lead wire and I don't like this.
I don't like it because I'm almost certain that pushing up to 150-180W can lead to disasters that are not visible by looking at the NTC temperature or even touching the heatsink plates

[orb]

Nella foto 4 potete vedere due piccoli dissipatori che ho messo sui diodi di protezione schottky e sulle resistenze di shunt.

Hello. I have the same pcb as yours. Thanks for the tip with the black heatsink on the shunt. I've bought two of "IRFP90N20D" too. I read on the internet that this mosfet works ok with dl24p.

[Filippo52]

Thanks for you replay.
Yes, the IRFP90N20D is one of the best replacements for the mosfet that the manufacturer supplies and has a little more power which is good, but I invite you to look at two things carefully:

1 - is the mosfet they sent you original?
I can give you two simple ways to check if it is with 95% certainty. The easiest thing is the weight. An original Mosfet such as IRFP90N20D must weigh about 5.5 g if it weighs 4.4 g it is a fake. Another thing you can try is RDS-ON; follow the diagram in the figure and if the value is close to that of the specifications or 0.03 ohm the mosfet will be original if instead it will be 0.1 ohm or worse it will be a fake.





if you need more details about this test ask and i will give you more information

2 - the second thing I want to tell you is that for this application the most important specification of the mosfet is that of the graph generally called "Maximum Safe Operating Area". I report the two images and you will see that the difference between the mosfet supplied by the manufacturer and the one you bought is less than what could appear from the maximum current and dissipated power

La prima immagine si riferisce al tuo nuovo Mosfet IRFP90N20D

[/url]

la seconda invece al mosfet che in questo periodo arriva già montato sulla board  ovvero IRFP264



The first thing you can notice is that both graphs lack the curve in "DC" and the closest possible one is the one at 10 msec. This is because both mosfets, unfortunately, are not designed for a linear application like this one of the DL24. But unfortunately we know that linear ones are difficult to find, especially in this period, and cost a lot, generally between €30 and €40.
We can therefore think of an imaginary line that has the same distance between the 1 msec line and the 10 msec one.

The operating area is the one at the bottom under the dotted line. As you can see in first image referred to IRFP90N20D at 1 volt you can apply max 30A (the line is the same for the various times) between 4 and 5 volts you have the maximum value that we can imagine equal to the maximum of the Mosfet of 90A, but at 10 volts you will have already dropped to 30A and at 100v you will only have 1A of current.

The same reasoning about the IRFP264 mosfet leads you to:
with one volt you have about 10A, at 3 volts you have a peak of about 25A, at 10V you will have about 70A (better than the IRFP90N20D) and at 100V you will also have only 1A for this

As you can see, the IRFP90N20D does better but not always and not as much as could be expected from the value of the maximum current and dissipated power.


to conclude, keep in mind that the heatsink that I see, even if it has a wider fan, I'm not sure can reach values of 180W; If you have experience with heatsinks used in PCs you can agree with my fear.

For the rest, if you stay under 100-max120W this is a really nice object

[Filippo52]

Dear ORB
I'll make an addition on the new mosfet you bought IRFP90N20D

in the following photo you see two of them next to each other: they are not the same.
One is a fake, the other is an original.



The one on the left is the fake and unfortunately it looks like your codes.

I invite you to use the weight in gr and the RDS-ON value to better understand
If you need any further clarification, I'm here
Hello

[orb]

@Filippo52, yes, probably you are right.
Both cost me around €8, so I must live with that fact.

I have bought also IRFP260N (looks like the genuine one, but who knows...):


I have kitchen scale only (1g resolution), here are the results:
IRFP260N: ~6g
IRFP90N20D: ~4-5g

Thanks for the info.

[Filippo52]

Dear Grab
even if your scale is approximate it is really very very difficult that an IRFP90N20D can weigh less than an IFFR260N (they are very similar to IRFP264), so it is probable that the IRFP260N you bought are genuine; the look of the numbering also looks right.
unfortunately, however, the weight of the IRFP90N20D is really low and together with the writing suggests that they are fake, even if the price you paid: 8 € is the right one for a genuine MOsfet of this type.
All you have left is the RDS-on proof to confirm; you should do this because a fake Mosfet can easily short-circuit and if you are working with a battery that can deliver very high currents you risk burning the board tracks and the components in series with the flux such as the double schottky protection diode and the two resistors shunts.

However, I'll give you another clue even if I don't know how valid it is: the back of the real mosfets that I have is different from the fakes that I have come across



in this photo you can see the back, the one on the left has the metal part with two notches and is the fake; the genuine one does not have these notches on the sides and has straight metallic sides.
Another small difference is the roughness of the plastic that covers the mosfet; in the fake ones it is more wrinkled

I don't know where you live (in which country) but I can give you component sites where you are sure to get genuine products

[Pukker]

Before buying components it is worth to check the company where you want to order
is a current/active member of ECIA (Electronic Components Industry Association)
in North America, ECSN (Electronic Component Supply Network)
in the United Kingdom, and CEDA (China Electronics Distributor Alliance) in China.
They are commited to buy from trusted sellers / manufactories.

You get your parts probably not for the low prices on Ali, Ebay ason,
but always getting headache from doubting "is this part genuine" is part of the low price.

BTW
High current load, high current battery, risk on short MOSfet : use an FUSE.

[orb]

However, I'll give you another clue even if I don't know how valid it is: the back of the real mosfets that I have is different from the fakes that I have come across



in this photo you can see the back, the one on the left has the metal part with two notches and is the fake; the genuine one does not have these notches on the sides and has straight metallic sides.
Another small difference is the roughness of the plastic that covers the mosfet; in the fake ones it is more wrinkled

Yes, exactly it is. IRFP260N seems to be fine. I use dl24p up to max 100W.

Dear Grab

Just like "grab this or grab that"?