Need replacement FET for obsoleted part

[zz28zz]

Here's the specs for the orig n-channel MOSFET:
https://www.digikey.com/product-detail/en/stmicroelectronics/STP50NE08/497-2787-5-ND/603812


Here's the one I think will work but would like someones opinion with more experience than myself:
https://www.digikey.com/product-detail/en/on-semiconductor/FQA46N15/FQA46N15FS-ND/1057222

I noticed the RDSON was a little higher on the new FET but not sure if that will make any noticeable difference in this application (Power FET for a car audio amp). The input capacitance is lower on the new FET but I think that's a good thing.

[amyk]

24m vs 42m Rds(on) is a huge difference, nearly double the resistance means nearly double the power dissipation and voltage drop.

PSMN017-80PS may be a suitable replacement, although the power rating is lower it has even lower Rds(on) which means lower power dissipation too.

[Mechatrommer]

Here's the specs for the orig n-channel MOSFET:
https://www.digikey.com/product-detail/en/stmicroelectronics/STP50NE08/497-2787-5-ND/603812

Here's the one I think will work but would like someones opinion with more experience than myself:
https://www.digikey.com/product-detail/en/on-semiconductor/FQA46N15/FQA46N15FS-ND/1057222

I noticed the RDSON was a little higher on the new FET but not sure if that will make any noticeable difference in this application (Power FET for a car audio amp). The input capacitance is lower on the new FET but I think that's a good thing.

i would not concern much about higher Rds(on) i suspect the unit will not make full use of 50A everytime (60W increase in the fet) so about increased heat, it may take, or it may not depending on analysis. but i'm more concern that those 2 are not footprint compatible (TO-3P vs TO-220). no objection if you can find or other can suggest better spec. just telling you on footprint.

[oPossum]

IRF3205

Very common in car audio amps from the past ~15 years, inexpensive for it's specs.

You may have to reduce the gate resistor values a bit due to higher gate charge (146 vs 85 nC)

[zz28zz]

Good catch on the package size.

I found a IRF3205 with a gate charge of 110NC@10V which is the same as the orig. RDSON is lower at 6.5mohm compared to orig at 24mohm.

https://www.digikey.com/product-detail/en/infineon-technologies/IRF3205ZPBF/IRF3205ZPBF-ND/666093

Do I need to worry abt too much current flowing thru the transformer? In other words, can too low of a RDSON cause damage to the windings?

[oPossum]

Lower R_dson is not a problem, higher can be.  The most important specs to match are gate capacitance and charge. Closest is best. Higher or lower can change gate rise/fall time enough to cause cross condition (shoot through).

[oPossum]

Good catch on the package size.

I found a IRF3205 with a gate charge of 110NC@10V which is the same as the orig. RDSON is lower at 6.5mohm compared to orig at 24mohm.

https://www.digikey.com/product-detail/en/infineon-technologies/IRF3205ZPBF/IRF3205ZPBF-ND/666093

Forgot about that one. The 'new and improved' IRF3205 - the IRF3205Z.

That would be perfect!

[Mechatrommer]

Higher or lower can change gate rise/fall time enough to cause cross condition (shoot through).

i think faster switching (lower gate capacitance is much more preferable as i think original circuit have put enough delay to avoid shoot through on slower switching time. faster switching also means more efficient circuit. ymmv.

[Zero999]

Higher or lower can change gate rise/fall time enough to cause cross condition (shoot through).

i think faster switching (lower gate capacitance is much more preferable as i think original circuit have put enough delay to avoid shoot through on slower switching time. faster switching also means more efficient circuit. ymmv.

The problem is when two transistors are used in a half bridge configuration, diodes and resistors are often added to avoid cross-conduction. The MOSFET's gate capacitance changes via the resistor (R2/R3), slowing down the turn on and discharges rapidly, via the diode (D1/D2), turning it off quickly. Changing the MOSFET's gate capacitance will alter the time constant and thus the delay, which can cause cross-conduction and large current spikes.

See simulation below. Note the huge current spikes drawn from the supply, when M2 is substituted with a device with a much lower gate charge.


Now this is making the assumption that the delays aren't implemented in the MOSFET driver circuit, in which case the change gate capacitance will make little difference. Imagine M1 and M2 are driven separately with the delays provided by the driver, rather D1 and R2 and D2 and R3 and the MOSFET gate capacitance.

[Mechatrommer]

Higher or lower can change gate rise/fall time enough to cause cross condition (shoot through).

i think faster switching (lower gate capacitance is much more preferable as i think original circuit have put enough delay to avoid shoot through on slower switching time. faster switching also means more efficient circuit. ymmv.

The problem is when two transistors are used in a half bridge configuration

you have a valid point there. and thats a nice half bridge config circuit there, a good reference for other project. but the OP mentioned car audio amplifier and i guess half bridge like you showed is not used. i guess it will be more like the circuit below.. 2 nfet pairs switched alternatingly on primary side (red colored) to produce Vac on the transformer output (secondary), in this case i believe fast charging and discharging will be preferable.

but i think if he want to change mosfet, he need to change all of them to the same part to avoid unecessary e-field (or h?) collapse inside the transformer, sort of equivalent to shoot through or shorted condition, similarly to your half-bridge setup, i suspect we need to change the upper p-fet to faster type as well to make them work again (or change gate resistance R2/R3?). ymmv.



btw: i'm quite dissapointed that you changed your username from "H" to "Z" you should deserve the "H"

[oPossum]

but i think if he want to change mosfet, he need to change all of them to the same part

Yes. They should all be replaced even if the exact part is used as a replacement.

[zz28zz]

Lots of great replies. Thx!
I'm not sure if mine is considered a half bridge or not, but here's a screenshot of the power side. Maybe that will help.

I'm a bit confused by the term gate charge. Is that the same as gate capacitance?

What exactly is cross-conduction? Is it when one primary winding starts flowing current while the other primary winding is still flowing current?

There are 6 FETs that control the primary windings and I'm planning on replacing all 6. Do they need to be from the same lot#?

[oPossum]

I'm not sure if mine is considered a half bridge or not, but here's a screenshot of the power side.

It is half bridge. Full bridge would have four [sets] of MOSFETs driving a single primary winding.

I'm a bit confused by the term gate charge. Is that the same as gate capacitance?

Not exactly the same. A MOSFET has capacitance between the gate and source. It isn't an ideal capacitor, so it is characterized in a few different ways. The gate capacitance is measured just like a capacitor. The gate charge is a measure of how much energy is stored in that non-ideal capacitor.

What exactly is cross-conduction? Is it when one primary winding starts flowing current while the other primary winding is still flowing current?

Yes. It occurs when both sets of MOSFETs are on at the same time causing substantial unwanted current flow through them.

There are 6 FETs that control the primary windings and I'm planning on replacing all 6. Do they need to be from the same lot#?

Some will say they must be exactly the same including lot code and date code. My recommendation is that they are the same part number from the same manufacturer.

[oPossum]

The totem pole driver transistors can fail when the power supply MOSFETs fail. Make sure the 2SD1919 and 2SB1277 transistors are healthy. Not shorted, or open. Sane beta / H_FE. Check the gate drive waveform before you install the new MOSFETs.

[Mechatrommer]

I'm not sure if mine is considered a half bridge or not, but here's a screenshot of the power side.

It is half bridge. Full bridge would have four [sets] of MOSFETs driving a single primary winding.

its the same as i posted earlier except left side to right and vice versa and driver is much more powerful totem pole bjt instead of directly to smps IC, i prefer to call them as low side switching (3 parallel mosfets in each pair/side/phase), where on the higher side is 12-15V and the transformers.

There are 6 FETs that control the primary windings and I'm planning on replacing all 6. Do they need to be from the same lot#?

changing them all with same lot/part number is the best, but if cost is an issue, you may use different part number (different spec), or newer parts with older parts (degraded spec) which are still working, but they will not be so ideal as they may not share load equally. depending on your situation, the mosfet that bears the most load may fails sooner, or it may not, so it depends. low side switching is much more forgiving than other fet/bjt configuration. ymmv.

[zz28zz]

I checked the gate waveform and it looks good to me. See attached. Should I try to put a load on it?

When ordering FETs from Digikey, can I make a comment that I want them all from the same lot #?
When I received the audio section FETs, I got all the same lot# for the low side, but 2 different lot #s for the high side. I always order 1 extra part when ordering so I wound up with 3 of one lot# and 4 of another. I put 3 ea on the high side of each channel. I'm assuming that was OK.

[Mechatrommer]

If you talk about same manufacturing date, i dont think thats necessary, it will be difficult otherwise. Even same batch sometime have slight performance variation. but if you can afford, why not? That should be better or lesser variation. Since cost is not an issue, you should stick to one type mosfet/part number and install all new, the older working fet you can keep as backup spare. When installing, you can install 1 mosfet 1st for each side and test at lower power sound volume. If everything ok, install all 3 mosfets for each side for full operation. current limit Psu may also help to see current consumption during idle. 0.3 to less than 1A idle current should be normal. Ymmv.

[oPossum]

Gate waveforms look good. Check them again once MOSFETs are installed. Should look about the same, possibly with a little ringing.

[zz28zz]

"Even same batch sometime have slight performance variation. but if you can afford, why not?" Are you saying requesting same batch# costs extra?

Don't have a current limiting p/s at the house.
Amp has three 30A fuses in parallel. I was planning on pulling them and installing a single 1 amp fuse for initial power-up.