Hi all,
Some time ago I build a tube amplifier. Since then amplifier has a “little” problem.
ONLY when I turn power ON I listen a BIG SCRATCH in my speakers (95dB).
Any comments or suggestions are very welcome (and sorry my English).
Tube amp details:
1 x 274B
2 x 300B (2 x 60mA)
2 x 5842 (2 x 10mA)
itot=140mA+/-
B+=430Vdc+/-
PS: I only have one 274B tube. I listen nothing when I replace it by a GZ37.
Katapum
Some time ago I build a tube amplifier. Since then amplifier has a “little” problem.
ONLY when I turn power ON I listen a BIG SCRATCH in my speakers (95dB).
Any comments or suggestions are very welcome (and sorry my English).
Tube amp details:
1 x 274B
2 x 300B (2 x 60mA)
2 x 5842 (2 x 10mA)
itot=140mA+/-
B+=430Vdc+/-
PS: I only have one 274B tube. I listen nothing when I replace it by a GZ37.
Katapum
Attachments
The 274B is a directly heated rectifier, and will warm up quickly. The GZ35 is indirectly heated, and will warm up slowly. (Your schematic show the symbol for indirctly heated diode, but on the GZ34, the heater and cathode are joined at pin 8. I think this is just a symbol inaccuracy).
So, just to confirm what I understand:
With 274B there is noise,
With GZ34 there is no noise..?
If the power supply is the cause:
I think the "scratch" noise may be because of the ringing caused by the 2 * 10H chokes and the capacitor combination 1u + 270n + 100n + 33n.
If this is the case, try adding 220K 2W resistor across each 10H choke.
Of course there could be many other reasons, especially some instability in the amplifier as it warms up. This is just my "gut feeling".
So, just to confirm what I understand:
With 274B there is noise,
With GZ34 there is no noise..?
If the power supply is the cause:
I think the "scratch" noise may be because of the ringing caused by the 2 * 10H chokes and the capacitor combination 1u + 270n + 100n + 33n.
If this is the case, try adding 220K 2W resistor across each 10H choke.
Of course there could be many other reasons, especially some instability in the amplifier as it warms up. This is just my "gut feeling".
The scratch you're hearing is probably from the rectifier arcing over internally at start-up, causing interference in the circuit via the 2 10H chokes. The 274B is a delicate tube, and quite prone to arc. Look at bright white sparkling spots on the filament during startup, this is usually a first sign of arcing, and a warning sign for more trouble. If you see this happening, quit using the tube in this application immediately, as the rectifier may fail in such a way that it takes other components with it.
I had the same problem with a 5U4G in a choke-input application (which should be easy on the rectifier during operation). The Russian version lasted longest, but eventually also started to sparkle and then arc. Directly heated rectifiers heat up so fast that the voltage rise caused currents to flow which saturated the input choke, which in turn caused the current to increase even more. With a GZ34 or GZ37 in the same position, measured current increases gently towards its nominal value as it heats up, and it comes nowhere near input choke saturation current (which, in my case, is about 500 mA) . As you have an input capacitor of only 1uF, your problem might be similar to mine, pushing the 2*10H chokes close to their saturation point.
There are two solutions: either a soft-start, that switches on full load after all tubes have warmed up properly, or a slow heating rectifier such as a member of the GZ family. The 274B reportedly sounds very nice, but it's too fragile to use without precautions.
I had the same problem with a 5U4G in a choke-input application (which should be easy on the rectifier during operation). The Russian version lasted longest, but eventually also started to sparkle and then arc. Directly heated rectifiers heat up so fast that the voltage rise caused currents to flow which saturated the input choke, which in turn caused the current to increase even more. With a GZ34 or GZ37 in the same position, measured current increases gently towards its nominal value as it heats up, and it comes nowhere near input choke saturation current (which, in my case, is about 500 mA) . As you have an input capacitor of only 1uF, your problem might be similar to mine, pushing the 2*10H chokes close to their saturation point.
There are two solutions: either a soft-start, that switches on full load after all tubes have warmed up properly, or a slow heating rectifier such as a member of the GZ family. The 274B reportedly sounds very nice, but it's too fragile to use without precautions.
Hi all,
I simulated all PSU with PSU2. PSU2 shows graphic without ringing problems.
274B looks ok at start up.
When I replaced tubes I didn’t change socket connections. Is GZ37 a DIRECT replacement to a 274B? If not how could I connect a 274B with two transformers (main and heater) without centred taps?
I fixed PSU schematic. Is it ok now?
Katapum
I simulated all PSU with PSU2. PSU2 shows graphic without ringing problems.
274B looks ok at start up.
When I replaced tubes I didn’t change socket connections. Is GZ37 a DIRECT replacement to a 274B? If not how could I connect a 274B with two transformers (main and heater) without centred taps?
I fixed PSU schematic. Is it ok now?
Katapum
Attachments
Hi,
If you have wired the rectifier as indicated in your circuit, that's the proper way to connect the GZ37, as it is an indirectly heated rectifier with the cathode tied to one leg of the filament at pin 8. This is where the HT should be taken off the rectifier.
Now I'm still wondering about this scratching sound. Ringing in the powersupply could cause some artefacts, but not scratching or cracking sound. What kind of 300B's do you use? And do they look OK when they start up? Because this scratch could also be caused by the filament of your 300B when warming up. As the GZ37 applies HT much later than the 274B, any scratching caused further downstream during warmup might be masked by using the GZ37. It is good to know the exact cause of your scratch sound and cure it, rather than simply suppress the symptoms by substituting a rectifier.
A simple test would be: remove the 274B from its socket, and turn the amp on. Wait for about 20 seconds and *gently but quickly* put the 274B in its socket. If you still hear the scratch, you have strong evidence of a rectifier problem. If you don't, you know you should look further downstream.
Jurgen
If you have wired the rectifier as indicated in your circuit, that's the proper way to connect the GZ37, as it is an indirectly heated rectifier with the cathode tied to one leg of the filament at pin 8. This is where the HT should be taken off the rectifier.
Now I'm still wondering about this scratching sound. Ringing in the powersupply could cause some artefacts, but not scratching or cracking sound. What kind of 300B's do you use? And do they look OK when they start up? Because this scratch could also be caused by the filament of your 300B when warming up. As the GZ37 applies HT much later than the 274B, any scratching caused further downstream during warmup might be masked by using the GZ37. It is good to know the exact cause of your scratch sound and cure it, rather than simply suppress the symptoms by substituting a rectifier.
A simple test would be: remove the 274B from its socket, and turn the amp on. Wait for about 20 seconds and *gently but quickly* put the 274B in its socket. If you still hear the scratch, you have strong evidence of a rectifier problem. If you don't, you know you should look further downstream.
Jurgen
Hi,
Have you tried plugging in the rectifier after the other tubes have warmed up? Please let us know the results, so we can really put a finger on this problem!
Jurgen
Plugging in a 274B in your case will not hurt anything, but the proper way to connect a directly heated rectifier is to take HT from a center tap on the filament winding. However the voltage drop across the tube is so large that this will not make a significant difference with the way things are connected now.
Have you tried plugging in the rectifier after the other tubes have warmed up? Please let us know the results, so we can really put a finger on this problem!
Jurgen
Hi,
First I replaced 300B Full Music by another pair (300B China??? Not Full Music) and I only listened a little scratch. Then I replaced 300B China by 300B Full Music again and I saw I big flash in 300B tubes. I tried it again and I saw the same. Then I replaced 274B by GZ37 and everything’s ok.
I almost lost 300 euro with this replacement.
PS: I followed your advice with a 300B China pair and I listened nothing.
Katapum
First I replaced 300B Full Music by another pair (300B China??? Not Full Music) and I only listened a little scratch. Then I replaced 300B China by 300B Full Music again and I saw I big flash in 300B tubes. I tried it again and I saw the same. Then I replaced 274B by GZ37 and everything’s ok.
I almost lost 300 euro with this replacement.
PS: I followed your advice with a 300B China pair and I listened nothing.
Katapum
Hi,
That is bad. I don't know the Full Music 300B's myself, but normally you would expect a 300B to be the tube that heats its filaments quicker than all other tubes in your amp, due to the lower filament mass. But it seems that your rectifier is no longer suspect. Flashes in tubes usually go together with cathode material flaking off so it is best that you don't repeat this experiment.
But it looks like your amp draws excessive cathode current at startup, so let's consider the power supply to be OK, and concentrate on the amp itself. Do you have a schematic, or a description of the topology?
Jurgen
That is bad. I don't know the Full Music 300B's myself, but normally you would expect a 300B to be the tube that heats its filaments quicker than all other tubes in your amp, due to the lower filament mass. But it seems that your rectifier is no longer suspect. Flashes in tubes usually go together with cathode material flaking off so it is best that you don't repeat this experiment.
But it looks like your amp draws excessive cathode current at startup, so let's consider the power supply to be OK, and concentrate on the amp itself. Do you have a schematic, or a description of the topology?
Jurgen
OK we now have 2 pages of a thread and a PSU schematic that shows a single vacuum diode shorting the HV winding, instead of the proper full wave rectifier or two diodes. As drawn the PSU will not work either with a 274B or GZ37 or ANY rectifier in this universe, because the last time I looked, Kirchoff was dead, which makes it very difficult to change his laws! I know it may be a minor quibble but it suggests that the original poster really has no idea what he is doing, and we would all like to think this is not the case.
In any case, the 274B should ideally be used with a center tapped heater supply to avoid a DC differential across the filament. I don't think this is very relevant for your case, though
Could the choke spikes be causing arc-over? Seems unlikely given the post rectifier caps, but then, you could always test this by replacing the choke with a suitable resistor. You would get more hum but at least you would be able to test if there is a scratching noise at startup.
In any case, the 274B should ideally be used with a center tapped heater supply to avoid a DC differential across the filament. I don't think this is very relevant for your case, though
Could the choke spikes be causing arc-over? Seems unlikely given the post rectifier caps, but then, you could always test this by replacing the choke with a suitable resistor. You would get more hum but at least you would be able to test if there is a scratching noise at startup.
Ilimzn: read post #11, the way the rectifier is shown here is a consequence of the limitations of the schematic capture software Katapum used. From his description, you can conclude that in reality the rectifier is wired correctly. A few posts later, he stated that it was the 300B that showed a flash. Which makes sense, because normally the last capacitor in the power supply would greatly suppress any audible effects of sparking/arcing further upstream, while the original description was about
Katapum, the power amp stage (5842 and 300B plus associated components) apparently draws excessive current when the power supply comes up quickly, otherwise your 300B would not have shown a flash. This makes me wonder about some component values and types:
- the grid leak resistor for the 300B
- the coupling capacitor trom the anode of the 5842 to the grid of the 300B
- the cathode resistor for the 300B
- the cathode bypass capacitor for the 300B
Also, a little more elaboration on the driver stage would be nice. Is it a CCS loaded 5842, where the CCS is placed where normally the anode resistor would be? I assume it is solid-state?
BTW, the stand-by implementation you drew is good, as even in stand-by a small current will flow, preventing the "sleeping sickness" problem.
Regards,
Jurgen
Katapum, the power amp stage (5842 and 300B plus associated components) apparently draws excessive current when the power supply comes up quickly, otherwise your 300B would not have shown a flash. This makes me wonder about some component values and types:
- the grid leak resistor for the 300B
- the coupling capacitor trom the anode of the 5842 to the grid of the 300B
- the cathode resistor for the 300B
- the cathode bypass capacitor for the 300B
Also, a little more elaboration on the driver stage would be nice. Is it a CCS loaded 5842, where the CCS is placed where normally the anode resistor would be? I assume it is solid-state?
BTW, the stand-by implementation you drew is good, as even in stand-by a small current will flow, preventing the "sleeping sickness" problem.
Regards,
Jurgen
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.