Inverter pre-charge thick film resistor

[gaminn]

Hi,
I'm looking for a flat thick film resistor, form factor approx as in the picture. Substrate dimensions are 34 mm height and 25 mm width. Approx 100 ohm resistance, should be able to pre-charge 1.3 mF capacitors to 320 V DC or 0.4 mF to 550 V DC.

These are available on eBay or AliExpress, used ones, removed from old boards and for 50 USD+ piece.

Is there any alternative? I need 10 to 20 pcs.

[Wolfram]

Does it specifically need to be thin film? For 60 - 70 J precharge like in your case, I would consider bulk ceramic or NTC. Something like a KOA HPC2 or HPC3 part should be sufficient here, as they have excellent pulse handling capabilities and overload margin compared to film and wirewound resistors.

PTC resistors can also be an option. This is a variation on bulk ceramic, where the resistance goes up with temperature, avoiding thermal runaway if the precharge gets stuck in the wrong state.

[gaminn]

No, it don't have to be thick film resistor. I'm mainly limited by dimensions (width, length and thickness)

That KOA HPC resistors have quite impressive one-pulse ratings https://www.koaglobal.com/-/media/Files/KOA_Global/EN/product/commonpdf/hpc.pdf. HPC3 has 2000 W 100 ms pulse rating and 100 W 1000 ms pulse rating. I can connect 2 resistors in series to double the voltage and power rating. However, the power rating for one pulse in their datasheet is for constant power pulse. My pulse is obviously not constant.

The power of the charging of 1.5mF capacitor via 100 ohm resistor by 50 Hz 320 Vac p-p is like this:



How would you translate this to constant power pulse?

[Zero999]

Film resistors are generally bad for peak power. Wire wound and carbon compositions resistors are much better at dealing with surges.

[gaminn]

Originally, these are the modules (the picture in my initial post) used in the inverter. It was designed 25 years ago approx, maybe they didn't have a better option that time. And I think these are thick film resistors.

I can probably create a little PCB with dimension like the original module and mount e.g. 2 KOA HPC3 resistors to it + some pins at the edge of the PCB to solder it to the main PCB.

[Wolfram]

Film resistors are generally bad for peak power. Wire wound and carbon compositions resistors are much better at dealing with surges.

Carbon composition resistors are mostly obsolete, and ceramic composition is the modern replacement. Fusing energy for a given power rating is something like an order of magnitude better than wirewound, on account of the larger mass of resistive material. I recently did destructive testing of precharge resistors for an application with very similar requirements, and the 2 W
ceramic composition resistors beat the 5 W wirewounds by a factor of 5 on average fusing energy.

[gaminn]

Hi,
I'm reopening this topic.

I tested KOA HPC3 ceramic resistors and unfortunately they are dangerous to be used as a replacement of the thick film resistors proviously used in the device. When there is a short circuit in the device (inverter), the thick film resistor simply explodes, opens the circuit and causes no harm. In the same situation, the ceramic resistor burns with a flame, probably thermally destroying components around and possibly causing a fire. There is no possibility to add a fuse in series with the resistor in my case.

I charge 700 uF to max. 500 Vac (= 700 Vdc). And it seems the suitable components to limit inrush current are PTC and thick film resistor. For PTC, I found https://www.vishay.com/docs/29165/ptcel_series.pdf, namely PTCEL17R251SBE, 250 ohm, 850 Vdc. I would connect 2 in parallel to provide approx 120 ohm charging resistance. Do you think this PTC can handle the situation or is there a better part?

For thick film resistors, I found this section on Mouser https://cz.mouser.com/c/passive-components/resistors/through-hole-resistors/thick-film-resistors/. However I have a difficulty to find auitable part that can handle the voltage and energy (approx 150 Joules of energy is delivered to the capacitor bank within approx 100 ms). There are many parts in TO-220 or similar case, however their single pulse overload rating doesn't look very promising. E.g. Bourns PWR220T-20 (20 W resistor) can handle like 10 Joules for 100 ms:


Not sure if the single pulse overload rating is calculated for a part mounted on a heatsink. I have no option to mount the part on a heatsink. Do you think there is a thick film resistor available on Mouser/Digikey that can handle 700uF/700Vdc? I can connect two in parallel if they are reasonably small.

Thanks

[ArdWar]

Continuous pulse withstanding, overload fusing, high voltage, benign failure mode. That's one hell of a requirement....

Maybe ohmite TP or TRH series if you can find them.

https://www.ohmite.com/assets/docs/res_tp.pdf
https://www.ohmite.com/assets/docs/res_trh.pdf

[gaminn]

I don't think these are excessive reuirements. The device is an 500 Vac inverter, with 700 uF DC bus capacitors, there must be thousands of products on the market with requirements like those mentioned. I wonder what is used inside these devices. Distributors of components don't have these Ohmite TRH and TP resistors in stock in large quantities. TRHs are also marked as obsolete.

Ohmite TRH, energy version, would do it easily I think. It has 100J pulse rating for 10 ms, 200J for 100ms. 100R version (TRHE01A100RJ2E) would charge for 70 ms (time constant), so it looks like this is perfect match. It is also available on Mouser. Until I measure the space for it on the PCB and find out that it is impossible to fit it on the PCB, as it is 15mm wide due to its integrated heatsink. That is another requirement - reasonable size.

Ohmite TP is 500 Vdc max, with this pulse rating: 100J: 100ms pulse with 100uF capacitor, not to exceed 1500 volts . I have 700 uF / 700 V, that is 150 J of energy, thus one Ohmite TP can't do it. What about two 50R in series? Can I just assume that two resistor in series can handle twice the energy? Then 2pcs 50R Ohmite TP could possibly handle it.

[T3sl4co1l]

Hi,
I'm reopening this topic.

I tested KOA HPC3 ceramic resistors and unfortunately they are dangerous to be used as a replacement of the thick film resistors proviously used in the device. When there is a short circuit in the device (inverter)

Ah, so it's incorrectly designed in general.  It's not the resistor's fault at all.  The design itself needs to be fixed.

It's sadly an all too common oversight, to employ a precharge circuit, and stop short of adding a voltage threshold and timeout to protect it.  Common result, resistor stuck on into shorted load, pffst-bang, whole thing's cacked.  Maybe just the resistor fails (fusible), maybe it melts and failure cascades.

A notable example I saw from a friend, a hi-fi audio amp used a wirewound resistor for this.  When the outputs failed shorted, the precharge dutifully turned on and stayed on, glowing red hot for, probably some minutes.  Now, this resistor was installed above the main PCB, so, it wasn't a pretty sight.  Smoke must've been pouring out of the poor thing before the user noticed it wasn't working.  I can only begin to imagine the smell... trying to clean that out is at least as bad as cigarette smell.   The board ended up serviceable, I think it was, no traces or components in the blackened area, but man... YUCK!

Tim

[gaminn]

I wouldn't say this is a bad design. The original thick film resistor explodes causing absolutely no other damage and danger.  Honestly, I don't think you can design it better, it is inherently safe. What if voltage threshold and timeout circuit fails? You still need a resistor that's not going to burn.

[T3sl4co1l]

Then replace with equal, fusible type resistor and pay the cost

[gaminn]

That is what I would like to do but I can't find many that can handle 150J charging. And this leads me to the question what others use when there are not many "safe" resistors products that are able to handle this.

What about PTCs, how they fail? E.g. PTCEL17R251SBE, it can handle 240 J, 850 V, it is small, I can use 2 in parallel to get the desired initial charging resistance. In case there is a shortcircuit in the inverter, they just get hot which I can handle somehow on the PCB. Will they ever burn if they fail, e.g. due to high voltage, exceeded number of cycles, exceeded consecutive cycles rating?

[Someone]

this leads me to the question what others use when there are not many "safe" resistors products that are able to handle this.

Automotive uses wire wound resistors bonded/coupled/integrated with a thermal fuse. But you're trying to find an exact fit for the existing part, which surprise surprise is only solvable with... the existing part.

[NiHaoMike]

Is your input supply AC or DC? If it's AC, maybe use a nondissipative current limiter like a capacitor or inductor?

[gaminn]

Automotive uses wire wound resistors bonded/coupled/integrated with a thermal fuse.

Can you post a link for that?

I don't need the original part. But the requirements are quite demanding regarding the size, energy, voltage. Of course the original part would be the best, but it is hard to get it.

The input is 3 phase up to 500 Vac. I have to physically place the inrush current limiter exactly where it was on the PCB, between the rectifier and the capacitor bank.

[Someone]

Automotive uses wire wound resistors bonded/coupled/integrated with a thermal fuse.

Can you post a link for that?

I don't need the original part. But the requirements are quite demanding regarding the size, energy, voltage. Of course the original part would be the best, but it is hard to get it.

The input is 3 phase up to 500 Vac. I have to physically place the inrush current limiter exactly where it was on the PCB, between the rectifier and the capacitor bank.

Its Automotive, they get parts custom made for them at volume pricing rather than trying to bodge in an off the shelf item.

[ArdWar]

Agreed that it will be much easier for you to consider your requirement separately. Although for "stuck conducting" part you may need to thermally couple your protection device to the resistor, yet preferably without compromising clearance rating when stuff blows.

[Wolfram]

Note that PTCs often have a significant voltage coefficient of resistance, decreasing with applied voltage.