Hi,
I've just finished building a Tubelab SSE and have a problem - it's mega quiet. I'm using Transcendar TT-312-OTs (SE, 5k, UL tap for 8 ohm speakers), Tung-sol KT120 power tubes, a Mullard CV4024 driver and a Sovtek rectifier. The amp is mounted on an aluminium plate. For now I'm driving a Mission 750LE speaker for testing (6 ohms not 8 but that shouldn't make a vast difference should it?). For a test input I'm just using a laptop playing mp3s. If I turn the volume pot all the way up I can hear music extremely faintly at the speaker (as in, I have to put my ear right up to the speaker to hear it) but there seems to be a huge gain problem somewhere. I checked the components very carefully and the sound is the same on both channels. I've tried both solid state and valve rectification and it doesn't make a difference.
As I understand it the KT120s are drop-in replacements for KT88s but can handle more power, and the CV4024 is equivalent to a 12AT7 so hopefully the tubes I've chosen are ok. I'd be very grateful for help troubleshooting this, as I'm stumped.
I've just finished building a Tubelab SSE and have a problem - it's mega quiet. I'm using Transcendar TT-312-OTs (SE, 5k, UL tap for 8 ohm speakers), Tung-sol KT120 power tubes, a Mullard CV4024 driver and a Sovtek rectifier. The amp is mounted on an aluminium plate. For now I'm driving a Mission 750LE speaker for testing (6 ohms not 8 but that shouldn't make a vast difference should it?). For a test input I'm just using a laptop playing mp3s. If I turn the volume pot all the way up I can hear music extremely faintly at the speaker (as in, I have to put my ear right up to the speaker to hear it) but there seems to be a huge gain problem somewhere. I checked the components very carefully and the sound is the same on both channels. I've tried both solid state and valve rectification and it doesn't make a difference.
As I understand it the KT120s are drop-in replacements for KT88s but can handle more power, and the CV4024 is equivalent to a 12AT7 so hopefully the tubes I've chosen are ok. I'd be very grateful for help troubleshooting this, as I'm stumped.
OK, I'll post some pictures tonight. Output binding posts are mounted through plastic collars and measure >20M to ground and to each other with the OPT secondary disconnected. I don't have a scope, just a multimeter, but I suppose I could use the laptop to generate a sine wave and measure RMS values.
A couple more questions in preparation for more fault finding tonight:
1) Should the DC resistance of the OPT secondary be 8 ohms, or is that just the AC impedance?
2) Can the tubes I'm using be operated upside-down? (This will make fault finding a lot easier but I gather there are some tubes where the orientation matters).
1) Should the DC resistance of the OPT secondary be 8 ohms, or is that just the AC impedance?
2) Can the tubes I'm using be operated upside-down? (This will make fault finding a lot easier but I gather there are some tubes where the orientation matters).
OK so I'm home now and have taken some photos and made some further measurements. First the photos:
http://www.adrianbaugh.org.uk/amp1.jpg http://www.adrianbaugh.org.uk/amp2.jpg
I guess the first thing is, can anyone see anything obviously wrong? (The disconnected wires near the front are from the volume pot, I chopped up a 3.5mm jack cable to make a very simple input from my laptop which is what is connected at the moment.)
And now the measurements:
I'm using a 440Hz sine wave generated on my laptop as input, measured at 1.0Vac using my multimeter.
Heater voltage: 6.50Vac. (The rectifier tube is unplugged for the moment, I'm using the diodes; the heater voltage to the rectifier socket looks okay at 5.3Vac.
B+ to ground measures 470Vdc.
I checked the OPTs too just in case they were dodgy: primary resistance was 410 ohms; secondary resistance was 1.3 ohms.
In the preamp, between the grid and anode measures 8.6V.
I'd be very grateful if anyone can point out a fault or suggest other measurements I could make to try to find the fault.
Not sure what else would be useful measurements
http://www.adrianbaugh.org.uk/amp1.jpg http://www.adrianbaugh.org.uk/amp2.jpg
I guess the first thing is, can anyone see anything obviously wrong? (The disconnected wires near the front are from the volume pot, I chopped up a 3.5mm jack cable to make a very simple input from my laptop which is what is connected at the moment.)
And now the measurements:
I'm using a 440Hz sine wave generated on my laptop as input, measured at 1.0Vac using my multimeter.
Heater voltage: 6.50Vac. (The rectifier tube is unplugged for the moment, I'm using the diodes; the heater voltage to the rectifier socket looks okay at 5.3Vac.
B+ to ground measures 470Vdc.
I checked the OPTs too just in case they were dodgy: primary resistance was 410 ohms; secondary resistance was 1.3 ohms.
In the preamp, between the grid and anode measures 8.6V.
I'd be very grateful if anyone can point out a fault or suggest other measurements I could make to try to find the fault.
Not sure what else would be useful measurements
I've taken the board out to check each component in turn, so here's another photo of the topside of the board: http://www.adrianbaugh.org.uk/amp3.jpg
I've swung through every resistor on the board and all the values are correct. All the diodes test correct and even though I've mounted the FREDs on the reverse of the board the dot on the package does line up correctly with the dot on the silkscreen layer. There is no visible sign of damage to any of the electrolytics and I've observed the NTC thermistor behaving correctly. All tube sockets line up correctly with the lug on the silkscreen.
I don't know how to test correct operation of the CCS ICs: when installing them I did handle them a fair bit before noticing my wrist strap wasn't plugged into the antistatic mat - are they particularly static-sensitive, and is there a way of checking rather than just replacing them on the offchance?
Next I need to check the tubes themselves. Are there measurements I can take with them disconnected that will indicate whether they are working? The KT120s are brand new so they should be okay (especially as the fault isn't channel-specific); the CV4024 is old stock and a potential point of failure common to both channels. I don't have a spare, but I could get my hands on a 12AU7 or a 12AX7 - albeit at reduced quality, would one of those suffice to test whether the fault disappears by swapping out the driver tube?
Thanks for all your help so far.
I've swung through every resistor on the board and all the values are correct. All the diodes test correct and even though I've mounted the FREDs on the reverse of the board the dot on the package does line up correctly with the dot on the silkscreen layer. There is no visible sign of damage to any of the electrolytics and I've observed the NTC thermistor behaving correctly. All tube sockets line up correctly with the lug on the silkscreen.
I don't know how to test correct operation of the CCS ICs: when installing them I did handle them a fair bit before noticing my wrist strap wasn't plugged into the antistatic mat - are they particularly static-sensitive, and is there a way of checking rather than just replacing them on the offchance?
Next I need to check the tubes themselves. Are there measurements I can take with them disconnected that will indicate whether they are working? The KT120s are brand new so they should be okay (especially as the fault isn't channel-specific); the CV4024 is old stock and a potential point of failure common to both channels. I don't have a spare, but I could get my hands on a 12AU7 or a 12AX7 - albeit at reduced quality, would one of those suffice to test whether the fault disappears by swapping out the driver tube?
Thanks for all your help so far.
I think it might be the CCS ICs: there is a faint smell of hot silicon (no smoke though) and the TO220 tab and heatsink are not at B+ as the semiconductors page of the assembly manual says they should be. I'll change them and see if that fixes things.
By the way, while I'm buying more parts, can anyone suggest a good (preferably UK based) source of 47uF 500V electrolytics? I couldn't find one before so there is a 68uF in there at the moment which is too big really.
By the way, while I'm buying more parts, can anyone suggest a good (preferably UK based) source of 47uF 500V electrolytics? I couldn't find one before so there is a 68uF in there at the moment which is too big really.
A 33u/500v will work if you can source one in the UK.... It's what I used. George states 22u-47u/500v can be used as C1 ...
Thanks, I'll see if I can find a 22 or 33u.
I found the main problem (hopefully)... I had installed the 10M45s the wrong way round. That's now corrected, with new devices, but the problem still exists - still very quiet, with a signal very faintly audible if I put my ear to the speaker. What other components am I likely to have damaged having those the wrong way round? I know the OPTs are good by the way, I checked them with a 12V 50Hz signal to the primary and got sound out of the speaker.
I found the main problem (hopefully)... I had installed the 10M45s the wrong way round. That's now corrected, with new devices, but the problem still exists - still very quiet, with a signal very faintly audible if I put my ear to the speaker. What other components am I likely to have damaged having those the wrong way round? I know the OPTs are good by the way, I checked them with a 12V 50Hz signal to the primary and got sound out of the speaker.
I've discovered something else. If I remove the driver tube (I've tried a CV4024, also a 12AU7 I had lying around), switch on and then insert the driver tube once everything else is warmed up, the volume level increases from nothing for a few seconds, becoming quite clear and audible (though not as loud as it should be given the source is turned right up) before falling away to the same extremely low level as before and staying there. Any suggestions, anyone?
Across R13 there is 0V dc, a few millivolts ac. (Same with R23.) DC voltages at the CCS pins are as follows: G - 0V, A - +475V as expected, K - 0V. When I touch the multimeter probe to the end of R13 there is loud hum audible from the speaker, it seems to be picking up mains hum via the probe.
So does this mean my second pair of CCS are defective? (or more likely I've broken them somehow - I hate handling semiconductors...)
Regards
Adrian.
With the driver tube out you should see the same voltage at A, K, and G. Make sure R10/20, R13/23, and R19/29 are not open and measure correctly. If they do, then it seems the CCSs are bad, though it's unusual for both to be bad unless they were purchased from a poor source or took way too much heat during soldering.
With the driver tube out, both CCS measure 0V at K and G and +475V at A. All the resistors measure correct so it looks like I got duff CCSs. They're difficult to get in the UK so I took a chance on eBay. I've now found a different supplier so I'll see how the new ones work out, they should arrive on Monday. Thanks for your help & I'll post an update once the new semiconductors are fitted.
The new parts have arrived. Before I fit them, a question about B+: when switching the amp on B+ increases to in excess of 500V for a few seconds (from memory about 540V) before settling back down to 475V. Could this be the cause of failure of the previous CCSs (and is it a problem for the electrolytics, which are 500V rated)?
I don't want to burn out another pair of CCSs if there's a power supply problem I need to sort out first!
I don't want to burn out another pair of CCSs if there's a power supply problem I need to sort out first!
The power supply is OK. It is normal for the B+ to surge on initial turn on, and your 500V electrolytics probably are rated for a 550V surge.
The voltages you stated around the CCS chips are wrong. This could be where the problem lies. You can test the circuit without the chips.
The 12AT7 will work with a plate load resistance in the 33 to 47K ohm range in place of the CCS. The performance will not be as good as a CCS (lower gain, higher distortion) but you can verify the rest of the circuit.
You can do this by jumpering the CCS and replacing the 10K 2 watt resistor with a something between 30K and 51K that is rated for 2 watts.
You can also leave the 10K in place and temporarily solder a resistor (20K to 33K) between the 10K and the 12AT7 plate (in place of the CCS chip).
The voltages you stated around the CCS chips are wrong. This could be where the problem lies. You can test the circuit without the chips.
The 12AT7 will work with a plate load resistance in the 33 to 47K ohm range in place of the CCS. The performance will not be as good as a CCS (lower gain, higher distortion) but you can verify the rest of the circuit.
You can do this by jumpering the CCS and replacing the 10K 2 watt resistor with a something between 30K and 51K that is rated for 2 watts.
You can also leave the 10K in place and temporarily solder a resistor (20K to 33K) between the 10K and the 12AT7 plate (in place of the CCS chip).
Success! (Progress, anyway...) I rigged up a 2W 30K resistor between the 10K and plate on one of the 12AT7s and that channel now works fine (and loud!) So... I'm guessing that narrows down the problem to a bad pair of CCSs, either through just chance or mishandling. I'll try fitting a new pair (carefully!) and see if that fixes things.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.