Frequency response issue solved
Ah well. Interesting lesson learned this evening.
Checked over the wiring (again) and couldn't see anything amiss, so started tracing the signal through the circuitry with the oscilloscope. At all points through the amp, I had the same frequency drop off, and eventually ended up back at the input socket where (you've guessed it) I still had the same issue. So, it obviously wasn't the amp. The only thing connected to it was the oscilloscope, which has it's own on-board signal generator which I normally use for testing. Although this has a variable output it does not go down to zero, so I feed the amp through a potentiometer and phono cable so I can vary the input level from zero up to several volts. I know the oscilloscope output is fairly consistent through the frequency range, but adding in the potentiometer and cable is having some effect causing the output to reduce with rising frequency. Would some kind person please put me out of my misery and explain what is happening.
Having determined that with the signal generator on minimum output it wouldn't overload the amp, I removed the potentiometer and additional cable and fed the amp with the signal direct from the oscilloscope. Success. All is good and I'm getting a nice flat frequency response well into the 50kHz region.
The testing has, however, highlighted a couple of points that I would appreciate some input on.
1. In checking the circuitry, it appears that the schematic generated by mullardel34 in the early days of this thread contains an error. The capacitor in parallel with the 47V zener on the grids of V2 is shown as 0.1uF, but on Ralph's original it is 1.0uF. I know some of the component values have been changed, but I wasn't aware that this was one of them. Am I better to stick with Ralph's value for this?
2. B2+ is down at around 280V, fundamentally as I am pulling around 13mA through the 3K resistor that has been put in the +ve arm of power supply. We use a couple of mA for the V4 bias voltage divider circuitry and I'm assuming 5 - 6mA per triode in V4 is fine, so it all seems OK, but would be interested to know if we need to increase B2+ a little by reducing the value of the 3K resistor.
3 The whistling only happens with the output stage connected to the HT, so it would seem that perhaps I have a dodgy output tube. Will see what happens once everything settles down.
Cheers,
Steve.
Ah well. Interesting lesson learned this evening.
Checked over the wiring (again) and couldn't see anything amiss, so started tracing the signal through the circuitry with the oscilloscope. At all points through the amp, I had the same frequency drop off, and eventually ended up back at the input socket where (you've guessed it) I still had the same issue. So, it obviously wasn't the amp. The only thing connected to it was the oscilloscope, which has it's own on-board signal generator which I normally use for testing. Although this has a variable output it does not go down to zero, so I feed the amp through a potentiometer and phono cable so I can vary the input level from zero up to several volts. I know the oscilloscope output is fairly consistent through the frequency range, but adding in the potentiometer and cable is having some effect causing the output to reduce with rising frequency. Would some kind person please put me out of my misery and explain what is happening.
Having determined that with the signal generator on minimum output it wouldn't overload the amp, I removed the potentiometer and additional cable and fed the amp with the signal direct from the oscilloscope. Success. All is good and I'm getting a nice flat frequency response well into the 50kHz region.
The testing has, however, highlighted a couple of points that I would appreciate some input on.
1. In checking the circuitry, it appears that the schematic generated by mullardel34 in the early days of this thread contains an error. The capacitor in parallel with the 47V zener on the grids of V2 is shown as 0.1uF, but on Ralph's original it is 1.0uF. I know some of the component values have been changed, but I wasn't aware that this was one of them. Am I better to stick with Ralph's value for this?
2. B2+ is down at around 280V, fundamentally as I am pulling around 13mA through the 3K resistor that has been put in the +ve arm of power supply. We use a couple of mA for the V4 bias voltage divider circuitry and I'm assuming 5 - 6mA per triode in V4 is fine, so it all seems OK, but would be interested to know if we need to increase B2+ a little by reducing the value of the 3K resistor.
3 The whistling only happens with the output stage connected to the HT, so it would seem that perhaps I have a dodgy output tube. Will see what happens once everything settles down.
Cheers,
Steve.
Interesting, I believe to have 0.1uf also
2. B2+ is down at around 280V, fundamentally as I am pulling around 13mA through the 3K resistor that has been put in the +ve arm of power supply. We use a couple of mA for the V4 bias voltage divider circuitry and I'm assuming
5 - 6mA per triode in V4 is fine, so it all seems OK, but would be interested to know if we need to increase B2+ a little by reducing the value of the 3K resistor.
/QUOTE]
Have 288-292 with the Atma-sphere XFA transformer and 1k resistors here
I have had (and some times still have...whistling, very sensitive speakers driven directly. However at some point I really noticed it to get a lot louder and was getting worried... obviously looking down on the tubes, I noticed spiderwebs between the tubes, (the d... critters have ADHD, special invasive breed that does not do web'sbut only pull strings.... )
anyway, gently removing the webs immediately removed the whistling or at least reduced it to a very faint noise.
I do not expect cob webs to be your issue, but it is a hint towards how sensitive the tubes can be to whatever the amp might be on top of if you are having it on the bench bottom up.
Thanks Demonkleaner. The only mods away from Ralph's original that I have implemented are the 5ohm resistors on the cathodes of the output tubes, and the increase of the input stage current to 5mA. Both mods seem to make sense based on my limited understanding of tube amps, and these at least were discussed in the thread.
Cheers,
Steve.
Hi Norgaard,
Thanks for that. Have added in another 3k resistor in parallel and now have 301V so much closer to the intended value.
Cheers,
Steve.
The zener bypass isn't critical. Its a good idea if you want to bypass for zener noise to set up the zener and its biasing resistor and then run a resistance to the grid- and have the cap to ground from there. Since the zener noise tends to be high frequencies the resistor need not be a high value but FWIW there is some cancellation occurring because the zener is tied to both grids of the tube.
285V is actually correct. You are dropping around 350volts across the tube already.
Well, they're alive and playing music.
All generally seems to be OK. Easily got both amps to zero dc on the output.
So the big question is - How many hours break in will I need?
For the first half our or so I genuinely thought I had a problem as there was something very strange going on in the mid-range, but that seemed to settle down quite quickly. They've now got around 4 hours on them, so starting to settle down, but still somewhat on the harsh side at the moment. That being said, they are already starting to show signs of their potential. Acoustic guitar is already superb, but I suspect they are going to be ruthless in their ability to show up any problems upstream
One problem I do have is 100Hz hum. It is very noticeable from the listening position so will need to look into that to see if it can be improved. I am currently using hybrid horn speakers, based on the Humble Homemade HiFi Calpamos, and these are around 98dB efficient so I suspect I will always get a little hum through the system. It is definitely the amps, as I still get it with a shorting plug on the input, so any thoughts and suggestions as to how to minimise it would be appreciated. As it is 100Hz, I am assuming it is from the rectified power supply rails.
Apart from that, all good. One amp does seem to generally have lower quiescent current in the output stage than the other, which I assume is something to do with the current through V4. Will let things settle down for a while and then look at tweaking the bias voltages.
Will report back in a week or two once I have a few more hours on them.
Cheers,
Steve.
All generally seems to be OK. Easily got both amps to zero dc on the output.
So the big question is - How many hours break in will I need?
For the first half our or so I genuinely thought I had a problem as there was something very strange going on in the mid-range, but that seemed to settle down quite quickly. They've now got around 4 hours on them, so starting to settle down, but still somewhat on the harsh side at the moment. That being said, they are already starting to show signs of their potential. Acoustic guitar is already superb, but I suspect they are going to be ruthless in their ability to show up any problems upstream
One problem I do have is 100Hz hum. It is very noticeable from the listening position so will need to look into that to see if it can be improved. I am currently using hybrid horn speakers, based on the Humble Homemade HiFi Calpamos, and these are around 98dB efficient so I suspect I will always get a little hum through the system. It is definitely the amps, as I still get it with a shorting plug on the input, so any thoughts and suggestions as to how to minimise it would be appreciated. As it is 100Hz, I am assuming it is from the rectified power supply rails.
Apart from that, all good. One amp does seem to generally have lower quiescent current in the output stage than the other, which I assume is something to do with the current through V4. Will let things settle down for a while and then look at tweaking the bias voltages.
Will report back in a week or two once I have a few more hours on them.
Cheers,
Steve.
Thanks Ralph. Have removed the additional 3k resistor so back to the 280V point
Hi Steve,
Great that you have them up and running and the music flows.
I have good - and bad? news. Mine are dead quiet, well with the ear right at 100dB efficient units, there is a slight sizzling, but no hum really. I have and have had, several solidstate designs that are significantly more noisy (and humming). The PSU's are as can be seen in the link earlier, non standard. But it is my impression that this is a trademark of the M60's
Cheers
Per
Great that you have them up and running and the music flows.
I have good - and bad? news. Mine are dead quiet, well with the ear right at 100dB efficient units, there is a slight sizzling, but no hum really. I have and have had, several solidstate designs that are significantly more noisy (and humming). The PSU's are as can be seen in the link earlier, non standard. But it is my impression that this is a trademark of the M60's
Cheers
Per
The first ten hours are not pretty.
Sometimes you can have a swept resonance problem between the inductance of the power transformer windings and the capacitance of the rectifiers. This can result in 'diode noise'. Keeping the power transformer leads to the rectifiers short can help reduce this noise.
Snubbing the rectifiers can also do the trick. A resistor- capacitor set in series bypassing each rectifier junction works whereas a single capacitor might do nothing or make it worse.
HEXFRED rectifiers can be used in the output. The trick with them is they are not forgiving of their surge current being exceeded and I'm kidding a bit- they don't have a surge current. So you simply have to get ones rated for enough current that they won't fail. We've had to snub them as well to achieve real silence, but this has paid off on horn speakers.
Also if you have horn speakers they might be so sensitive that you simply don't need the gain. If the speakers are over 101dB or so this is likely the case. You can reduce the gain by not using the tube at the top of the cascode voltage amplifier- simply remove it and jumper the cathode connections to the plate connections such that you have a simple differential amplifier instead.
That will knock down a good bit of noise and the sound quality will not suffer.
Hi Guys,
Thanks for the info.
Norgaards description of the noise as 'sizzling' is probably more appropriate than to say I have some hum, so I assume this is all pointing to diode noise as suggested.
I have hexfreds on the +/-300V supply as per mullardel34's schematics, and standard 4pin bridge rectifiers (25A 400V) for the 2 output stage power supplies. Capacitance values are very much as spec'd.
So if I can pick your brains a little further -
Is the input stage or output stage the more likely culprit (or both)
What value of capacitor and resistor would you suggest for the snubbers, and again do I need them on all the power supplies.
Thanks for your continued support.
Cheers,
Steve.
Thanks for the info.
Norgaards description of the noise as 'sizzling' is probably more appropriate than to say I have some hum, so I assume this is all pointing to diode noise as suggested.
I have hexfreds on the +/-300V supply as per mullardel34's schematics, and standard 4pin bridge rectifiers (25A 400V) for the 2 output stage power supplies. Capacitance values are very much as spec'd.
So if I can pick your brains a little further -
Is the input stage or output stage the more likely culprit (or both)
What value of capacitor and resistor would you suggest for the snubbers, and again do I need them on all the power supplies.
Thanks for your continued support.
Cheers,
Steve.
Just adding a little more to this.
I like Ralph's idea of using Hexfreds on the output stages, so have found some 8A 600V units at RS Components. So have ordered some of them and will get them soldered in at the weekend. Will also rig up another dummy load tonight to add to the one I use for testing, so can then put the SACD player on repeat and get some hours on the amps without waking the neighbours.
Will let you all know how things go.
Cheers,
Steve.
I like Ralph's idea of using Hexfreds on the output stages, so have found some 8A 600V units at RS Components. So have ordered some of them and will get them soldered in at the weekend. Will also rig up another dummy load tonight to add to the one I use for testing, so can then put the SACD player on repeat and get some hours on the amps without waking the neighbours.
Will let you all know how things go.
Cheers,
Steve.
I used these: VBE17-06NO7 IXYS | Mouser
I think I mounted them pins down through a pcb with a small heatsink to be safe.
Zero issues to date.
I think I mounted them pins down through a pcb with a small heatsink to be safe.
Zero issues to date.
OK, amps have been left running for 2 full days playing music into a dummy 8 ohm load, so have had a good soak test. Came through it fine with no issues. One amp needed the zero offset re-adjusting, but apart from that all is well.
Have switched out the output stage bridge rectifiers for Hexfreds, but still have got some buzzing / sizzling through the speakers so am going to have to investigate a little further and look at trying some snubbers. Fortunately it's not too obtrusive once the music is playing, so am not overly concerned at this stage.
They now have around 50 hrs on them and sound quality has improved significantly. They are now very listenable, and still improving. Midrange and treble are excellent. Bass is a little soft at the moment. I'm a huge fan of the prog rock band Yes, and Chris Squire's bass just doesn't quite have the growl it should have. Hopefully this will improve as more hours are put on the amps. I know Ralph has specific views on output impedances, but I do wonder if the high output impedance of the amps is contributing to this issue. I suppose that the zero autoformers are an option, but it just seems extremely illogical to me to build a superb OTL amp and then stick a transformer between it and the speakers.
I would be interested to hear other peoples views on this.
Overall, I'm extremely pleased with the amps. A few teething issues to resolve, but that's one of the challenges of DIY stuff. Thanks again to Ralph for making the schematic available to the DIY community
Cheers,
Steve.
Have switched out the output stage bridge rectifiers for Hexfreds, but still have got some buzzing / sizzling through the speakers so am going to have to investigate a little further and look at trying some snubbers. Fortunately it's not too obtrusive once the music is playing, so am not overly concerned at this stage.
They now have around 50 hrs on them and sound quality has improved significantly. They are now very listenable, and still improving. Midrange and treble are excellent. Bass is a little soft at the moment. I'm a huge fan of the prog rock band Yes, and Chris Squire's bass just doesn't quite have the growl it should have. Hopefully this will improve as more hours are put on the amps. I know Ralph has specific views on output impedances, but I do wonder if the high output impedance of the amps is contributing to this issue. I suppose that the zero autoformers are an option, but it just seems extremely illogical to me to build a superb OTL amp and then stick a transformer between it and the speakers.
I would be interested to hear other peoples views on this.
Overall, I'm extremely pleased with the amps. A few teething issues to resolve, but that's one of the challenges of DIY stuff. Thanks again to Ralph for making the schematic available to the DIY community
Cheers,
Steve.
Thanks Demonkleaner. Have dome the mods with mixed results, and things unfortunately are getting worse rather then better.
Measuring the voltages at the cathodes of V3 and V4, I get -90V at the cathode of V3A (from V1), -210V at the cathode of V3B and -70V at the cathodes of V4
The heaters for V1 and V2 are in parallel off one winding of the transformer and have been tied to ground, using 2, 220 ohm resistors across the heaters to effectively create a centre tap for the transformer winding. The earth connection is connected between the two resistors.
The heaters for V3 and V4 are also in parallel off a second winding, and again through 2 220 ohm resistors. These have been tied to a voltage divider off B-(2) at -140V. This puts all 4 6SN7's within the +/-100V limit between heaters and cathodes.
Whilst this has reduced the buzzing slightly, I now have a lot of spitting and popping from the amp. More worryingly, it now squeals when turned off.
The spitting and popping is similar to a problem I had with a Transcendent Sound phono stage, and this was cured by putting a 1k resistor in series at the input socket. Tried the same thing here. No change.
But here is the real clue (potentially) to the problem.
I have a fairly long run of cable from the input socket to V1, around 300mm (or 12 inches for you guys across the pond) due to the layout of the chassis. I'm just using a phono socket, not running balanced with an XLR). This is done with a twisted pair of wires, soldered from the input socket to 2 solder lugs just above V1. The 4.75k resistor is then soldered between the solder lug and the tube socket. The 100k grid leak is connected from the tube socket to the star ground.
I have an earth wire running from the star ground to the solder lug where the earth wire from the input socket is soldered. Now here's the interesting bit. If I remove the earth wire from the solder lug and solder it directly to the input socket, the spitting and popping gets significantly worse. DC wise, the circuit is identical, as all I am doing is moving an earth wire from one end of another piece of wire to the other, but this now has oscillation written all over it.
So if anyone has any suggestions as to where to start sorting this out, I would be extremely grateful.
Cheers,
Steve.