Hello. I'm new here and I'm looking for some design help. I'm designing an amp and in the phase inverter (typ long tail pair) is where I'm introducing negative feedback from the output transformer. Think very typical fender/marshal arrangement.
For some reason, when I connect the NFB (I'm using a connector to make it easy to switch transformer taps) it appears to be grounding out the power tubes. I'm not positive that that's what's happening but I notice the blue glowing cloud between the plate and cathode of each power tube disappears which to me says the tubes aren't conducting but I could be wrong about that as well.
Everything appears to be grounded correctly. My negative grid bias is pretty conservative right now putting the tube into about 60% of max dissipation. These are two EL34s in push-pull.
The amp works perfectly without the feedback. It obviously isn't going to have the correct sound I'm looking for without it tho.
Another possibility (instead of the tubes being grounded out) I'm considering is that the secondary of the output transformer is the opposite phase of what I think it is thus making it positive feedback. This feels right because when powered up with the NFB connected, the electronics hum at an increasing pitch til it reaches a maximum pitch which takes just a few seconds. Note this is the electronics humming, not the speakers. This happens without speakers connected. I don't have a scope, so I'm not sure on this.
If this is the case, what ways are acceptable to reverse the polarity of that signal? Should I change how I connect the secondary of the output transformer? For example, make the 8ohm leg the new common, the 16ohm will be the new 8ohm, the 4ohm would stay the same, and the common would be the 8ohm? This would change the polarity of the 4 and 8 ohm legs allowing me to use one them as NFB. The 16 would still be the same polarity but I just won't use that one.
Many thanks for any help
For some reason, when I connect the NFB (I'm using a connector to make it easy to switch transformer taps) it appears to be grounding out the power tubes. I'm not positive that that's what's happening but I notice the blue glowing cloud between the plate and cathode of each power tube disappears which to me says the tubes aren't conducting but I could be wrong about that as well.
Everything appears to be grounded correctly. My negative grid bias is pretty conservative right now putting the tube into about 60% of max dissipation. These are two EL34s in push-pull.
The amp works perfectly without the feedback. It obviously isn't going to have the correct sound I'm looking for without it tho.
Another possibility (instead of the tubes being grounded out) I'm considering is that the secondary of the output transformer is the opposite phase of what I think it is thus making it positive feedback. This feels right because when powered up with the NFB connected, the electronics hum at an increasing pitch til it reaches a maximum pitch which takes just a few seconds. Note this is the electronics humming, not the speakers. This happens without speakers connected. I don't have a scope, so I'm not sure on this.
If this is the case, what ways are acceptable to reverse the polarity of that signal? Should I change how I connect the secondary of the output transformer? For example, make the 8ohm leg the new common, the 16ohm will be the new 8ohm, the 4ohm would stay the same, and the common would be the 8ohm? This would change the polarity of the 4 and 8 ohm legs allowing me to use one them as NFB. The 16 would still be the same polarity but I just won't use that one.
Many thanks for any help
Please tell us you are powering up with a dummy load attached instead of the speakers. Otherwise you are likely to damage the OPT.
To reverse OPT polarity you either swap the primary or the secondary.
If this is about an instrument amp it ought to be in the instruments and amps forum.
To reverse OPT polarity you either swap the primary or the secondary.
If this is about an instrument amp it ought to be in the instruments and amps forum.
Yessir. Dummy 8ohm load to keep down any noise which may arrise from powering up a newly built amp.
It's going into an instrument amp but this is more about the tube circuit design. Not sure how to classify.
It's going into an instrument amp but this is more about the tube circuit design. Not sure how to classify.
Threads on valve instrument amps belong in the instruments and amps forum. It says so at the top of the page.
Without a circuit diagram we can't help much.
Without a circuit diagram we can't help much.
Not sure of the frequency as I don't have a scope but it's quite audible and doesn't sound low enough to be 60hz or 120hz. My schematics are hand drawn and at home. I could try to snap a pic tonight and put them up. The output stage is effectively identical to common marshall output stages with two el34s. Something like a JCM800 50W version if that helps.
I'm more interested in knowing how experienced designers would expect an amp to behave if the phase of their output transformer was backwards thus giving positive feedback. Would you expect it to take what little noise was available and build it up over a few seconds? Also, I'm wondering what would cause the electronics to make an audible sound? The transformers are brand new so I can't imagine the laminations vibrating/humming.
I don't think you can use GNFB injected on the cathode of the LTP. Both cathodes are common so you would have positive FB on the non-inverted half. Also you must keep in mind that coming straight off the secondary with FB you will have a low impedance path to ground through the tranny that will then reduce the LTP cathode resistance since it is paralleled. You must have a FB resistor in the circuit or else the driver bias will then be out of whack.
Both Fender and Marshall did this. The FB is connected to the grid that is not fed by the preamp, and some of the FB also goes to the cathodes. A knob affects the amount fed to the cathodes (the Presence control).
No just swap the two plate leads. Leave the secondary alone.
Somewhat aside - but I am worried about this glib reference about connecting up gNFB and simply reversing it if you get an oscillation.
That might well be the last output you will get from that experiment. Such testing with full feedback the wrong way round has been known to take out output transformers at least.
When one is unsure of the phase, leave off the NFB, use a small input signal and then apply feedback with a resistor some >100x the circuit value. It should cause no oscillation, but will show either a small increase/decrease in output signal, depending on the phase - that way you will know which way round is negative FB.
That might well be the last output you will get from that experiment. Such testing with full feedback the wrong way round has been known to take out output transformers at least.
When one is unsure of the phase, leave off the NFB, use a small input signal and then apply feedback with a resistor some >100x the circuit value. It should cause no oscillation, but will show either a small increase/decrease in output signal, depending on the phase - that way you will know which way round is negative FB.
Thanks for the advice. I'll try swapping the plate leads of the output transformer. For some reason I didn't think that would work since it's center tapped and the current always enters the center tap. I just assumed that it would always create the same phase. I actually had to draw out the transformer and analyze the two different connection options for the plate leads. Now that I've done that I see that it will indeed change the output phase. I'll never make that mistake again.
I think I'm going to bite the bullet and buy a two channel scope so that I can just look at the two points I'm wanting to connect to confirm that they're truly out of phase before hand. I've checked the output transformer and it seems to have not been ruined by the previous erroneous connection.
Thanks so much for all the help. I'll be sure to update.
Have you check if the 0 (zero) wire of secondary is connected to ground?
Only in this case you can send back (by resistor) a FB signal from one end ( 4 or 8 ohm tap).
This FB signal, normally, goes to the cathode of the first stage and not o the LTp and must be in phase; as written , you can swap the anode connection of the power tubes to get the right configuration.
Normally if you get the positive feedback you hear a incremental noise from the output transformers, when you power on the amp. You have time to switch off!!!
Ciao
Walter
Only in this case you can send back (by resistor) a FB signal from one end ( 4 or 8 ohm tap).
This FB signal, normally, goes to the cathode of the first stage and not o the LTp and must be in phase; as written , you can swap the anode connection of the power tubes to get the right configuration.
Normally if you get the positive feedback you hear a incremental noise from the output transformers, when you power on the amp. You have time to switch off!!!
Ciao
Walter
As previously suggestesd since this is a guitar amp I will move to I & A per forum policy.
Many DMMs will detect an ultrasonic oscillation. Simply set the DMM to AC Volts and connect across the load. Normal noise would measure something below 100mV, an oscillation will read many volts. Disconnecting the feedback should stop the oscillation as a double check.
Since you have a multi tapped output transformer, you shouldn't reverse the secondary. You can reverse the primary at the output tubes, the grids of the output tubes or the output of the phase inverter. Only reverse the phase at one place.
The feedback connection to the bottom of the tail resistor makes that resistor look more like a constant current source. This helps balance the phase inverter, which is of questionable value in a guitar amp.
Since you have a multi tapped output transformer, you shouldn't reverse the secondary. You can reverse the primary at the output tubes, the grids of the output tubes or the output of the phase inverter. Only reverse the phase at one place.
The feedback connection to the bottom of the tail resistor makes that resistor look more like a constant current source. This helps balance the phase inverter, which is of questionable value in a guitar amp.
Without an actual schemo, it's hard to tell much. If it's anything like this (attached) I can tell exactly what happens, since Mr. Murphy initially wired it for positive feedback. When advancing that feedback knob from zero, for awhile, nothing happened at all. Then, a small level, sinusoidal oscillation at about 1100Hz appeared: the phase neutral frequency. Adding more feedback first made it increase in amplitude. Then it became a slightly distorted sine.
Enough feedback caused it to oscillate like a Royer: distorted suqare waves at about 20Hz that brought the 8R test load resistor to red heat in about a minute. Here, the frequency determined by the volts/turns ratio of the OPT.
That meant switching the plate connections to the primary side of the OPT to reverse the phase of the feedback from positive to negative. Then, no more problems.
With speeks connected, the results would be immediate, and deafening.
Attachments
- Status
- Not open for further replies.