• Disclaimer: This Vendor's Forum is a paid-for commercial area. Unlike the rest of diyAudio, the Vendor has complete control of what may or may not be posted in this forum. If you wish to discuss technical matters outside the bounds of what is permitted by the Vendor, please use the non-commercial areas of diyAudio to do so.

Mosfet breakdown

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
This wiring is true for constant current sources (like LM336) but isn't true for MOSFET's. You will need to rewire your circuit to use P channel MOSFET's as CCS or BJT's. You need to generate a voltage of about 3-4V between gate and source (depending of the polarity of the MOS transistor) to turn it on (Class A biased), and then adjust source resistor to get drain current you need. Power in it will be directly I * E across the mos, then multiply it by the thermal resistance from die to air, and you will get temperature rise in the MOS without the heatsink. If this rise is too high, use a proper heatsink needed to reduce temperature to normal (< 70 deg Celsius) value.

But, why not to use the other half of the 'AT7 as a SRPP?

I don't have enough knowledge to understand what you are say :eek: I just learned enough to understand and build this amp.... at least i thought :rolleyes:

Each half of the AT7 amplifies a channel... at least i think it does.

My problem is with hardware... when it played it showed ~320V at mosfet imput and right now i have the same voltage as B+ :mad: The 12AT7 plate voltage was ~200V and now it's residual :mad:

Once again i will do a thorough look .... i hope i willl find something wrong

Thank you Osvaldo
 
Try this MOSFET CCS: as I say, the diode must be tried at between 2 to 4V depending of the MOSFET's VGSth, and then the source resistor selected to give the current you need from it. R2 must be gives say, 1mA to zener diode, and the voltage of VDD-Vz/Iz. Discard gate current. Power in the MOSFET is VDD times Io.
 

Attachments

  • FCC.GIF
    FCC.GIF
    5.2 KB · Views: 152
Try this MOSFET CCS: as I say, the diode must be tried at between 2 to 4V depending of the MOSFET's VGSth, and then the source resistor selected to give the current you need from it. R2 must be gives say, 1mA to zener diode, and the voltage of VDD-Vz/Iz. Discard gate current. Power in the MOSFET is VDD times Io.

Thank's Osvaldo for the time and help given :up:
I will study it and find diodes.
 
I think it would be much easier just to install the 10M45S. It will "just work".

IXCP10M45S IXYS | IXCP10M45S-ND | DigiKey

If getting them from Digikey is cost prohibitive, perhaps you can find a proxy? Peter Millett was going to sell them for his popular DCPP amp, which also uses them.

Surely, if the circuit has been designed for it, but the guy has the liberty of try other solution(s) as all of us.
 
I think it would be much easier just to install the 10M45S. It will "just work".
...

I already buy them on the auction site and have them installed.

Are you using FETs in place of the CCS? If the CCS is doing it's job, the voltage will vary as the CCS maintains a constant current. What is the voltage *across* R13/23 and R14/24?

I don't get readings across these resistors... 0 VDC

B+ Rises until ~470/480 VDC and decrease and stables around 430/440 VDC

Thanks :up:
 
I already buy them on the auction site and have them installed.

OK, so you are no longer trying to use a FET here? Good.

I don't get readings across these resistors... 0 VDC

Zero volts across all four resistors? That would mean that the CCS and/or tube are not passing any current at all. So let's review:

You originally had DN2540s here, but were experiencing part failures so you switched back to the 10M45S. You have a 330ohm resistor in there now, right? I assume the resistors still measure the correct resistance.

Looking at your other voltage measurements and assuming the resistors and PCB are good, it seems that the 10M45S is an open circuit right now. You have B+ on R14/24 and a few volts on R13/23.

So the question is, what is taking out your FETs/CCSs? Perhaps you have a shorted input tube? I assume that the heat sinks/hardware are not touching your chassis anywhere, right? The tab on the CCS is connected to the "anode" of the CCS, so it will have B+ on it.

What are you using for an input tube? Do you have a spare to try?

B+ Rises until ~470/480 VDC and decrease and stables around 430/440 VDC

That sounds normal...just the output tubes warming up and conducting.
 
Surely, if the circuit has been designed for it, but the guy has the liberty of try other solution(s) as all of us.

Of course, but we are still trying to diagnose an unknown problem with his amplifier. It is a reliable design, so some other factor is at play. Introducing more complexity at this stage is probably not going to help. ;)
 
You will need to rewire your circuit.....But, why not to use the other half of the 'AT7 as a SRPP?

There is no need to rewire or redesign the circuit, it works great the way it is designed. Both halves of the 12AT7 are already used.

It's a proven design with others running without fan aid. I'm doing something wrong.....This is how i assemble them:

That looks wrong.

It is not clear if you are putting the CCS chips on the top of the board or the bottom of the board. Either way your picture does not match either drawing in the assembly manual. If you can't find the assembly error, please post a picture of the CCS chip mounted on the board, and we will figure it out.
 
OK, so you are no longer trying to use a FET here? Good.



Zero volts across all four resistors? That would mean that the CCS and/or tube are not passing any current at all. So let's review:

You originally had DN2540s here, but were experiencing part failures so you switched back to the 10M45S. You have a 330ohm resistor in there now, right? I assume the resistors still measure the correct resistance.

Yes, checked all of them


Looking at your other voltage measurements and assuming the resistors and PCB are good, it seems that the 10M45S is an open circuit right now. You have B+ on R14/24 and a few volts on R13/23.

So the question is, what is taking out your FETs/CCSs? Perhaps you have a shorted input tube? I assume that the heat sinks/hardware are not touching your chassis anywhere, right? The tab on the CCS is connected to the "anode" of the CCS, so it will have B+ on it.

What are you using for an input tube? Do you have a spare to try?

No, the sinks don't touch anything, they are inverted assembled.

I have a few 12AT7 and tried some with the same results.... I checked the condutivity in the board and everything seems right :S

I must be doing something really stupid and assume it as right :warped:

I'm posting some photos, please take a look too


Thank you Russ
 
...

That looks wrong.

It is not clear if you are putting the CCS chips on the top of the board or the bottom of the board. Either way your picture does not match either drawing in the assembly manual. If you can't find the assembly error, please post a picture of the CCS chip mounted on the board, and we will figure it out.

Thank you George :up:

I looked and looked several times and can't find anything wrong :confused:

As i said before, within this solder/unsolder i damaged the tracks... to make sure a perfect conductivity i replaced them with wire as you can see. It wasn't rocket science but one never know and i might screwed up :eek:

An externally hosted image should be here but it was not working when we last tested it.


An externally hosted image should be here but it was not working when we last tested it.


The tracks conections between the mosfet and R13/23 and R19/29 are in the other side of the board.
The track between C21 and R23 and R29 was replaced with a wire. At R23, a wire pass through the board to the other side to establish connection with the driver tube. Checked conductivity within these points.

An externally hosted image should be here but it was not working when we last tested it.


An externally hosted image should be here but it was not working when we last tested it.


Replace tracks with wire, conductivity checked.


An externally hosted image should be here but it was not working when we last tested it.


10M45 bended. I assembled them on the lower side of the board.

Can you spot anything wrong?

Yepp, my board is a mess and the suposed cleaning liquid didn't help too :rolleyes:


Thank's
Rui
 
Hmmm...something is wrong with those measurements. The plates (pins 1 and 6) match the DC offset on one of the heater pins (pin 5). Are you sure those measurements are correct?

Perhaps it's because the heater has AC and not DC on it and it's confusing your meter. Or, you may have a short between one of your plates and the heaters. The heater supply is floated to around 20-30VDC (depending on B+) by the divider comprised of R4 and R3. So you should see this DC offset on all of the heater pins (4, 5, and 9). If you measure across the heater (pins 4-9 or 5-9) with your meter set to AC, you should see the 6.3VAC heater supply.

A short between the plates and the heater might explain the short live of your CCS's. I could be the tube, but I would inspect you soldering around the socket on both sides of the PCB very carefully.
 
Hmmm...something is wrong with those measurements. The plates (pins 1 and 6) match the DC offset on one of the heater pins (pin 5). Are you sure those measurements are correct?

You are right, they are inverted... they are correct now:

Pin 9 - ~29 VDC
Pin 8 - 0 VDC
Pin 7 - (-0.5)VDC
Pin 6 - (-0.8) VDC
Pin 5 - ~29.5 VDC
Pin 4 - = Pin5
Pin 3 - 0 VDC
Pin 2 = (-0,5)VDC
Pin 1 - (-0.8) VDC

Perhaps it's because the heater has AC and not DC on it and it's confusing your meter. Or, you may have a short between one of your plates and the heaters. The heater supply is floated to around 20-30VDC (depending on B+) by the divider comprised of R4 and R3. So you should see this DC offset on all of the heater pins (4, 5, and 9). If you measure across the heater (pins 4-9 or 5-9) with your meter set to AC, you should see the 6.3VAC heater supply.

Measured across them and got ~6.1VAC

A short between the plates and the heater might explain the short live of your CCS's. I could be the tube, but I would inspect you soldering around the socket on both sides of the PCB very carefully.

Nothing wrong here then?


Thank you
 
Looks fine. There is nothing on the plates, which is expected given that we know that your CCS's are likely not working. Have you already tried replacing them? Maybe we should try replacing the CCS with a resistor to make sure everything else is in order before blowing up any more silicone.

Yes, already had with the same result :S

I'm open to all suggestions, I need help although ;) I have a few spare resistors and i can buy some in the hardware store.


One more thing... i just take some readings at R13 and initially the voltage rise until ~40VDC and then lowers to ~0VDC. I only have one multimeter but the time interval matches the one B+ takes to rise until 450/480 VDC.... this brings something?
 
Last edited:
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.