Check very well the VS resistors (4 x 1ohm or 2.2 or 4.7ohm) they can be phiscally ok, but not electrically.
Change them in all cases.
Check also all gate and pulldown resistors. Sometime some resistors can change their value (minimum change can be catastrophic when you apply the power.)
Change them in all cases.
Check also all gate and pulldown resistors. Sometime some resistors can change their value (minimum change can be catastrophic when you apply the power.)
Checked High Side on each IC. It has the same amplitude like the low side, but the High side is glichting on the time base... Should it be like that? Low side stays stable on the time base.
Yes, readjusted a little, now its ok. I have a good signal on low and high on each bank. Will recheck the VS resistors . Something else i can check before mount the fets?
tested with low power, everything looks good, after a few seconds, two banks high side fets blown again....
Were you monitoring the temperature?
Did they heat up really quickly?
Did it destroy the driver ICs again?
What make/model is your scope?
Did they heat up really quickly?
Did it destroy the driver ICs again?
What make/model is your scope?
I dont monitored the temperature. I left it out of the heatsink, let it run for half a minute,everything looks fine, Amplitude was also ok. Put it back in the heatsink, checked, that there is no short. Applyed power, after the current draw goes down, when the IRS start swichting, directly the fets exploded.
From measurment, it seems, that the drivers still ok.
Scope is UNI-T UTD1025DL
Outside the heatsink, the fets stay cold. I tested with my finger... Maybe a intermittend problem.
From measurment, it seems, that the drivers still ok.
Scope is UNI-T UTD1025DL
Outside the heatsink, the fets stay cold. I tested with my finger... Maybe a intermittend problem.
What 'amplitude' was OK?
Does the rail voltage build fully before the outputs start switching?
For problems like this, using a low-voltage rail supply can sometimes be useful. Are you familiar with that?
It's still possible that the inductor could be intermittently shorted.
Does the rail voltage build fully before the outputs start switching?
For problems like this, using a low-voltage rail supply can sometimes be useful. Are you familiar with that?
It's still possible that the inductor could be intermittently shorted.
Yes, i know that low voltage supply. I had a full amplitude at the drain. Like it should be.
Ok, i removed the mosfets. One fets per each bank is gone. So 4 pieces. One driver is damaged, the other 3 drivers still working. I have to order new drivers. I think also, that it could be the inductor or something intermittend....I dont know, if its worth to continue...
And yes, the rail was fully up, before switching starts.
Ok, i removed the mosfets. One fets per each bank is gone. So 4 pieces. One driver is damaged, the other 3 drivers still working. I have to order new drivers. I think also, that it could be the inductor or something intermittend....I dont know, if its worth to continue...
And yes, the rail was fully up, before switching starts.
With low voltage you can use much more robust FETs to see if you can find the problem.
Just because it had rail-rail oscillation, that doesn't mean that the drive signal was good. A weak drive signal can make the FETs run hot, even though it's getting them to switch on and off.
When only one 21844 fails, is it always the same one?
Just because it had rail-rail oscillation, that doesn't mean that the drive signal was good. A weak drive signal can make the FETs run hot, even though it's getting them to switch on and off.
When only one 21844 fails, is it always the same one?
Yes, its always the same driver.
In the beginning, when i tested with only two fets, all four drivers are gone...
But the damaged drivers are mostly the two on the right side of the drive board.
In the beginning, when i tested with only two fets, all four drivers are gone...
But the damaged drivers are mostly the two on the right side of the drive board.
There are two diodes on the drive board, and the caps. Can you give me the correct value or better a link for mouser... I will order some parts.
And this clips for the fets, to clamp them to the heatsink, also i cant find them anywhere. Do they have a specific name?
And this clips for the fets, to clamp them to the heatsink, also i cant find them anywhere. Do they have a specific name?
Last edited:
The capacitors vary from 4.7uF to about 10uF. 5.6uF and 6.8uF are likely the most common.
The diodes have to be fast recovery and capable of easily withstanding the full rail-rail voltage. The ER1G is common.
How many clips do you need? Someone may have a few that they can send you but I have seen no source that could supply them in bulk.
The following clips can be used for TO-220 parts. In a pinch, you can use 2 on the larger semiconductors. They are not reusable. They do not apply enough force to flatten component that are not soldered in parallel to the heatsink. They are not perfect but have worked reliably to get me through repairs where clips were missing or broken.
4525 Aavid Thermalloy - Thermal Management - Distributors, Price Comparison, and Datasheets | Octopart component search
The diodes have to be fast recovery and capable of easily withstanding the full rail-rail voltage. The ER1G is common.
How many clips do you need? Someone may have a few that they can send you but I have seen no source that could supply them in bulk.
The following clips can be used for TO-220 parts. In a pinch, you can use 2 on the larger semiconductors. They are not reusable. They do not apply enough force to flatten component that are not soldered in parallel to the heatsink. They are not perfect but have worked reliably to get me through repairs where clips were missing or broken.
4525 Aavid Thermalloy - Thermal Management - Distributors, Price Comparison, and Datasheets | Octopart component search
Ok, and the small signal diodes on the drive board just standard?
For the Metall clips i need 12 pcs, for two fets each.
Later i want to start with the low voltage test. Should i take out all railcaps for this?
You mentioned, i can use stronger fets? I still have alot irfp064n, should they work for testing?
For the Metall clips i need 12 pcs, for two fets each.
Later i want to start with the low voltage test. Should i take out all railcaps for this?
You mentioned, i can use stronger fets? I still have alot irfp064n, should they work for testing?