add-on: by "comparing voltages"  I meant comparing them in working and non-working channels.

    Quote:   Do you have a multimeter? If you don't then I definitely recommend buying one (HarborFreight had them for $3.50). With multimeter in hand I'd do the following: since nothing is burning you can test the  circuit as is. Draw a table, name the key points of the circuit...

most probably you either short-cut the mosfets or fried them resulting in a complete open mosfet. You can check if the filaments are still operable by testing the resistance between the corresponding pins (#3 and #4). For the future tests you can first use a powerful resistor instead of...

  Quote: There is a simple way to get rid of this problem: diode protection. The diodes should be switched in parallel to earphones and each channel will need either 2 of them (to kill both forward and reverse spikes) or 4 of them (to double the threshold above which the spikes/sounds...

if the filaments don't glow that means either burnt filaments (measure them with Ohmmeter when the tubes are out of the sockets) or burnt mosfets or closed mosfets.

all voltages are measured with respect to the ground.   The grid voltage in this connection should be = 0 (there's no voltage source at the potentiometer and the grid current must be negligible).

the correct way to test the circuit is to measure the voltages in key points. The cathode voltage should be >0 and less than a couple of volts, the grid voltage in this scheme must be equal to zero, and the anode voltage should be less than 48 V but greater than 40 (my guess, I don't have...

it's interesting... I wonder how it can be that 150K resistor switched in series with 50K pot doesn't help. It's 1:4 attenuation.

there are simpler ways to cope with high gain/9o'clock pot problem: add a resistor in series with your pot and you'll easily go beyond 9 o'clock

2 revolink24:   thermal material is __necessary__ to insulate your mosfets from the ground. Otherwise you'll have to insulate the radiators that might be tricky. Check the list of items provided by Peter Millet on the first page. I recall there were two mounting sets that included thin...

in principle, nothing bad should happen given that you've put the output caps. To be on a safe side you can start from the output: measure the DC voltage on the output jack. If there's no DC component, switch to AC measurement mode and check if the amp doesn't generate by itself when the...

First of all, do NOT measure the resistance when the system is powered. You can fry your multimeter.   Second, I'd suggested to use multimeter in voltage measuring mode to know for sure where is "+" and "-". Naturally, this measurement should be performed when the PSU works. Regardless...

check the resistance of the heater. The one that is working shines too bright for me - normally they should glow. I suspect the heater was burnt. If that's the case, look for the reason. Otherwise you'll fry the second tube.

I meant the resistance of the whole circuit when you unplug the power jack and put the ohmmeter leads to "+" and "-" of your circuit.

wait a second - you mean that the voltage was not equal to 48V when BOTH tubes were unplugged? And at the same time the unloaded PSU shows 48V? I'm a bit confused. Can you measure the resistance of the unpowered circuit without tubes? It might be that the PSU has gone crazy so it can't power...

that's exactly what I meant. If the PSU can power one channel and if you do the same test for the second channel (just to make sure they are identical and there are no surprises to expect) then you know for sure that you need either a more powerful PSU or separate power switches for the...

add-on - if the mosfets die they can do it in two ways: always open and completely burnt when there's no conductivity at all. In the second case there will be no load so the PSU voltage will not drop. If we're dealing with the first case then they will not act as current limiters and the PSU...

It looks like your power supply units do not like the cold heaters. I mean they do not withstand the initial current. There are two solutions: 1) buy a more powerful PSU. 2) Install a switch (or two switches) that will allow you first to start one channel, then the second one.    

Antoine,   the easiest and safest way to do is to use the multimeter (seriously). Take a power jack, plug the power connector, measure the voltages, write them down, and so on. I always double-check that even if I know the pinouts. Plug the earphones to the output jack and measure the...

next time when you mount your new MOSFETs, try to "reverse-engineer" your circuit. I mean, draw a circuit diagram looking just at your connections. If you get the original scheme, congratulations. Then you can take the ohmmeter and check for the shortcuts. If you don't find any, power the...

it's not clear from the pictures if the the MOSFETs are electrically insulated from the radiator/ground. If they are not then you ground the sink that for this MOSFET means short-cutting the power supply. May be all the troubles started with that, the MOSFETs were fried and everything got...

2 nightarya: what do you mean by "nothing"? No voltage at either of the points? What do you measure when the power supply unit is unplugged? (I mean, does it work at all? - it sounds stupid but I faced with many cases including the faulty connector that looked absolutely normal and new)...

Did you measure the voltages in the key circuit points? What were the results? I'd suggest disconnecting everything except the heaters and to make them working properly. After that one can move further.

Hi everybody and many thanks for this thread.   Here’s my build. I modified the circuit a bit to use 6J6 tubes that are still abundant and available at ~$2 a piece. The modification is straightforward: one will need two 39 Ohm 10W resistors (mouser part# 71-CP10-K-39) and two 150 Ohm...

Hi everybody and many thanks to the author of the thread for simple yet effective design.   I tried to reproduce the scheme but it's next to impossible to find the 19J6 tubes nowadays. So, I ended up with slight modification of the original circuit diagram to fit it to 6J6 tubes that are...