Hoffman Amplifiers Tube Amplifier Forum
Amp Stuff => Tube Amp Building - Tweaks - Repairs => Topic started by: angelodp on July 27, 2024, 07:33:58 pm
-
Hi, per an old post I have converted a Bogen 110. I will include the schematic from Dummyload, albeit a bit modified from that suggestion. The amp now sounds pretty good and goes into a nice OD breakup. I had to use a couple of Plate Load Split resistor setups and two lower Mu tubes in V1 and V2. My question is relating to the NFB circuit in this amp as it is. I am not familiar with what this part of the circuit is doing. I know what a NFB circuit typically does. When I bypass the NFB (disconnect the 12k wire to 16ohm tap) the sound is quite unpleasant. I have implemented numerous variable NFB pots in Fender style amps and so I am wondering how I can achieve that level of control on this amp.
-
The 2.7k resistor is your bias for v3. When you say disconnect it, are you lifting a leg of the resistor? If so, the amp won't work properly. You'd have to disconnect the nfb wire itself.
-
I am disconnecting the green wire that feeds into the 16ohm tap from the 12k R - (correcting the above post) to hear what a NFB defeat (like with a Fender) sounds like and it's a mess. That whole NFB network is foreign to me. I have managed the preamp end to sound pretty good and the power section is more or less untouched.
-
Assuming everything is working as designed, wired properly;
pull V1, temp in this, leave NFB DIS-connected
does this get you closer to happy sound?
-
Hi, is your suggestion to simply jumper over with NFB disconnected, or to also disconnect at the red x marks. I tried just jumping with the NFB disconnected and not getting a good result.
-
RED X is a cut, you're taking out 2 gain stages, my guess is you don't have a scope so troubleshooting
quite unpleasant.
becomes troublesome.
-
Ok yes, it does work nicely. The jumping no, but the direct link bypassing the stage yes. I suppose I can now try a variable NFB? Thanks for the input.
-
Just an observation - It can be hard to give up gain stages.:icon_biggrin: But with too many gain stages, folks end up doing what you did, substituting lower MU preamp tubes and adjusting resistors and caps to lower the overall gain. One of the problems with that is that every tube, cap, resistor. etc. adds some level of noise. Now with a more appropriate number of gain stages, you not only can play with adjustable FB, but you can also be free to try different preamp tubes, caps, and resistors to shape the tone.
Those Bogen amps are high quality so you should end up with a fine amp. And take a look at Tubenit's Carolina Special that is in the other Bogen 110 thread. His projects are supafine IMO.
-
Appreciate the observation. Yes I have that Carolina project on tap. My thought was to do a bit with the amp as is and get a feel for where its at. I may just go that Carolina route.... it will require gutting it pretty much. The Carolina project does make great use of all the tubes and only requires an additional reverb T.
Thanks A
-
The R32 resistor is forming a voltage divider with the cathode resistor of the triode that the NFB is feeding. I'm not sure exactly what the fix for your amp is, but I recently guitarified a Bogen/Challenger CHA33 that had similar "quasi-fixed biasing." It was pointed out to me by more experienced folk that this was likely done to allow these amps to run with severely degraded tubes, increasing the reliability of the units. I ended up re-configuring a significant portion of the power supply to get rid of those voltage dividers.
-
Ok thanks for pointing that out. Not sure how to remedy that. I think I may prefer the tone I had going with the NFB in and the cascading gains. I am using an airbrake to boil things down. I may temp in a switch to go back and forth with the V1 and another for the NFB. I had originally intended to go with the Carolina idea and may still.
-
Hi again, here is a clean copy of the original amp. My question is for a more clear understanding as to the NFB circuit and how it operates and what is the effect on the PI circuit. Any info on the nature of this part of the circuit and what its function is would be greatly appreciated.
-
I do not know the electronic details of that NFB circuit - other than what you state you already know about how NFB works. But, what I do know (or think I know) is that audio amps tend to have a lot more NFB than guitar amps. With an audio amp the designers want to avoid distortion, more so than guitar amps. They inject more NFB into the circuit to tame it. So when you disconnect the NFB, its a big (often nasty) change. For most guitar amps the change is subtle. That schematic is also not drawn the way that most guitar amp schematics are drawn, but bringing NFB into both cathodes of the PI is not unusual.
So if you want to keep all those gain stages - not advisable IMO - you need that high level of NFB that a small resistor allows. If you cut back on the gain stages, you can increase the value of the feedback resister and add a pot if you wish for adjustable feedback.
All of that is from my limited level of knowledge - and all could be wrong. But if so, smahter folks will jump in and help us both out. :icon_biggrin:
-
Thanks, I implemented a pot to control the NFB level (50k). That is still way too much defeat and I am going to try a 10k pot and go with the existing NFB and variable level with the 10k up to 22k. The way the power stage is designed does do a very good job of cleaning up any noise in the amp.
-
Were you able to measure the nominal dc circuit voltages in your amp, as shown on the schematic? Were you able to confirm the output stage 6V6's have about the same cathode currents, eg. with a bias probe or by inserting say a 10R in series with each cathode ? Are you using a DC voltage meter with a 10Megohm input resistance ? These tests are best done with 12k NFB resistor disconnected, and an 8 or 16 ohm resistor across the related output transformer tap, to minimise the chance you have some unobserved instability going on in the background.
The output circuitry that includes the GNFB loop is somewhat complicated, as it includes a 3M3-1M divider to drop the dc voltage to the PI stage input, and that also includes some output stage common cathode signal, and has a 4n7F across the 3M3. The PI stage also locally feeds back its output signal via the 100k to the first stage's global feedback summing node. The output stage has a capacitor bypass on its common cathode, and some common screen voltage bleed through 100k. The input stage also has a 39k-270pF high frequency step filter across the 470k anode load. That all makes for a lot of RC filter corner frequencies to appreciate along with the normal PI to output stage coupling CR, and the unknown response of the output transformer to low and high frequencies. Sorting those affects out would require some serious gain-phase plotting and stability margin testing that may be beyond you. If your dc levels are nominal, and you have a signal generator and ACV meter, then you could measure the output Vac with, and without, feedback resistor connected, using a low signal level with suitably low frequency that your meter can measure, and from those two output voltages then calculate the mid-band feedback dB level being applied as stock. That dB level may indicate how 'hi-fi' they were trying to make the amp (eg. say about 12dB), compared to a guitar amp that may use say up to 6dB.
-
Thank you for your thorough response. Yes I have a good meter, Extech 350, that meets the 10M requirement. The nominal DC voltages were good and the bias was in inline albeit a cursory set of measurements. I will do a more complete and accurate review of DC voltages and bias readings. This Bogen amp certainly is well made. My aim was to better understand this circuit and not necessarily try to 'fix' a tried and true idea beyond my capability. At the moment I do have an amp that plays nicely with useable tone perhaps louder than I require. My Airbrake is helping with that. My question now is whether to venture forward with Tubenit's Carolina amp schematic. I have gutted a number of amps and rebuilt them so I know I can take that on. I appreciate your taking the time to explain this area of the circuit and I now have a much better sense of what is going on with the NFB design.
Many Thanks Ange
-
A bit of an update. I decided to increase the ratio of the first split plate load resistor to 70k/30k. Now with the reduced ac signal the NFB pot has a useful range. The amp is quiet and both the Treble and Bass control are effective. With the airbrake in the mix I was able to go back to 12AY7 in V1 and then 12AX7 for the remaining preamp and PI tubes. Been playing it and enjoying the tones found within.
Would you think that adding grid stoppers 2.2k and screen resistors 470ohm or 1k would be prudent on this conversion?
Thanks agin for the insights and support
Ange
-
The advantage of adding grid-stoppers is you can easily short them out - for comparisons.
The higher the signal level coming into a stage, the more likely a grid stopper can alleviate harshness from an overdriven input.
Only V1A and V3A have a grid stopper, so all others may get some benefit, depending on how hard you can push each valve's input. V3B grid stopper may need to be quite high in value (eg. 100k to 1M) to have an effect.
Have you fused the HT secondary CT and/or inserted a 1N4007 protection diode to each 5Y3 anode ?
-
Can you expand on the comparison idea please?
I was thinking in terms of grid stoppers and screen resistors for the 6V6 tubes.
I can do the protection diodes for sure. Although this amp has the usual convention of red and yellow wires reversed. Its 4&6 that the diodes want to
end up with the PT wires and other end of diodes on 3&5. Wonder why they did not follow the usual color convention.
Fuse the HT CT - would that be a 1-amp fuse? Simply insert between T and Ground.
The other grid stoppers on pre-amp are they necessary?
Thanks A
The advantage of adding grid-stoppers is you can easily short them out - for comparisons.
The higher the signal level coming into a stage, the more likely a grid stopper can alleviate harshness from an overdriven input.
Only V1A and V3A have a grid stopper, so all others may get some benefit, depending on how hard you can push each valve's input. V3B grid stopper may need to be quite high in value (eg. 100k to 1M) to have an effect.
Have you fused the HT secondary CT and/or inserted a 1N4007 protection diode to each 5Y3 anode ?
-
Comparisons with/without the resistor. Likely needs cranked playing and maxed vol pot settings to notice the difference. Depending on your pups, or input signal, preamp stages could well be pushed into grid-current territory. Merlin has a good write up on cranking the concertina PI stage, which may then get uglier being in the GNFB loop.
-
Would you think that screen grid resistors and grid stoppers on the 6V6 a good move. I have the protection diodes in.
-
The output stage screen supply has 22uF smoothing - that value can be played with to suit your preference. Similarly, individual screen stoppers may change the character of overdrive due to sagged screen voltages, which is when each screen separately can draw a lot more current than during normal class A and AB operation. The sag is the instantaneous voltage drop across the stopper, as well as voltage droop and rebound on the 22uF cap. Some make the stopper an indestructible resistor, and others just make its power rating survivable for typical cranked operation but also a poor mans fuse if the particular screen fails as a short to cathode or whatever.
Similarly for input grid stoppers (for any stage) where the grid could be pushed towards and into grid current conduction, and hence the stopper resistor affects how abruptly that can happen.
Different guitar amp makers seem to have their preference for values of grid and screen stoppers, and screen filtering. Whereas hi-fi generally tries to make the stopper values as insignificant as practical.
-
Hi Tim, thanks again for the in-depth response. I have added the 1.5k grid stoppers on two small turrets and after an initial bit of sputtering (re-flowed the joints), its in there and I can pick up some benefit at the full tilt volume going into clipping OD. The NFB pot (50k) has some value at the lower level, so adding to that 12k up to about 20k is useful then the tone gets a bit muddy. Now this amp has the original tubes and my guess is that the 6V6 tubes might be in need of a refresher. Bias is a tad cold at 10watts per tube.
Thanks again for the insights.
Ange
-
Tim on the 100k screen bleed R's....will the insertion of the 470ohm screen grid resistors between the 100k bleed and the screen pin on the tube be correct? The 100k stay? Yes?
-
A new concern has arisen. What is the best way to implement a MV before the PI on this schematic?
-
A new concern has arisen. What is the best way to implement a MV before the PI on this schematic?
Before the grid of the stage that picks up the global NFB.
-
Ok, great - so a 1 meg audio should do it? In place of the 33k? Correction, had the wrong stage.
-
I have a 500k pot as a var. R in place of that 33k and the attenuation of signal is very small?? I need some help understanding how this could be more elective. Do I need to wire the pot so that the third leg is dumping signal to ground?
Ok so yes, I had is wired incorrectly, now does act as a volume pot, but not the desired effect. I was hoping that I could crank the previous stages and with the initial volume then retain some breakup but lower volume with the master. So fail.... Is there a better way? Post PI? Is that effective with this style PI?
-
Perhaps place a gain pot after the first triode, instead of the split anode load which is cutting down gain anyway. You could also modify the Volume control to work directly from the anode, rather than again being cut down by its split anode load connection.
Modifying the following tone and output/feedback circuitry is not so straightforward, including trying to alter how the volume of a circuit with global feedback around it works such as playing with pre or post PI circuitry.
Depending on your input signal level, an overdriven signal into V2A can be played with by adding a grid-stopper to V2A, and changing V2 to another 12A.7 tube, and going back to how V1A and V1B are managed for low and high frequency content via modifying their coupling cap and cathode bypass cap values.
Wrt the output stage 100k screen resistor to cathode, the schematic seems to show two such resistors in parallel. If so then that effective 50k may be just to lower the idle screen voltage a bit. I don't think it is intended to bring up the cathode by 2V during start-up, as that seems marginal, or that it aims to reduce start-up voltage stress on the 22uF screen decoupling cap as that cap is part of a can that is exposed to full B+. Using individual screen stoppers is a separate aspect of operation and so inserting them doesn't per se change how the 2x 100k are operating.
-
I have a 500k pot as a var. R in place of that 33k and the attenuation of signal is very small?? I need some help understanding how this could be more elective. Do I need to wire the pot so that the third leg is dumping signal to ground?
Ok so yes, I had is wired incorrectly, now does act as a volume pot, but not the desired effect. I was hoping that I could crank the previous stages and with the initial volume then retain some breakup but lower volume with the master. So fail.... Is there a better way? Post PI? Is that effective with this style PI?
Not knowing how you have the volume pot wired, try running the wiper into the 33k grid stopper, with the bottom leg to ground, and the top leg into the previous stage coupling cap.
-
Here's how I implemented the MV. It is somewhat effective but I realized that my anode split loads are stealing too much gain now, so I will try to bring gain slightly back. I also want to try a 1uf instead of the 4.7uf cathode bypass cap at V1. I am hoping to get the bass frequencies reduced and a tighter feel.
Thanks for the input.
-
Full schematic at present
-
Here's how I implemented the MV. It is somewhat effective but I realized that my anode split loads are stealing too much gain now, so I will try to bring gain slightly back. I also want to try a 1uf instead of the 4.7uf cathode bypass cap at V1. I am hoping to get the bass frequencies reduced and a tighter feel.
Thanks for the input.
Yes, maybe 4 tubes isn't enough gain.
You could try deleting the 68k resistor, and see what happens.
Your treble control also influences gain from the ground side.
-
So R23 is shown on the schematic with dash lines and the juncture shows an X....not sure what that means.
The 68k.
-
Bypassing the 68k kills the treble response and lowers gain.
-
So R23 is shown on the schematic with dash lines and the juncture shows an X....not sure what that means. The 68k.
Perhaps best to read the manual:
https://audiocircuit.dk/downloads/bogen/Bogen-DB110-int-sm.pdf (https://audiocircuit.dk/downloads/bogen/Bogen-DB110-int-sm.pdf)
-
Yes I see that, not used in some versions. So I have that 68k and bypassing it is not "good" for my purposes. I did implement a cold clipper at V3b and getting some 'roar' going. That along with the 1uf/1.5k at V1A is more to my liking.
So I ask - what's up with the 68k not being used in some amps? How was that effective?
Sorry if I am lacking in scientific know how. I am experimenting 'by ear' and learning some useful lessons here.
-
The 'X' path provides a local feedback signal. The output signal from V2b is being fed back to its input, and the output is loaded by just the treble control pot tap resistance (unknown), or that unknown R in series with 68k. That output signal, or its attenuated level (due to the 68k) is connected to the bass control circuitry. If the bass pot wiper was at bottom, then the feedback signal would mix with the dry input signal at the 470k summing point. As the bass pot wiper moves to the top, C14 allows more higher frequency signal than lower frequency, and so bass is not reduced by feedback as much, and so bass output increases. Hopefully I got that right. Perhaps worthwhile looking at Presence control feedback and how that works as a learning curve.
-
Thanks for that analysis. Not such a simple little amp at all. BTW bridging the 68k R22, seems to be a bit of a RAW link. That bridge (could be a 100k pot) has opened up a lot of options including a more Fendery glassiness in the treble control. The MV now has more range into a breakup visa vie the reg volume. I don't want to start drilling a bunch of holes in the front-plate for these mods just yet. I also tried a cold clipper at 10k for V2B, but reverted back to a standard 2.1k with a 4.7uf in parallel. Swapping listening, trying a variety of combinations.
-
Corrected, I do not have 4 pre-amp tubes so Princeton will not work. Back to one tube reverb. Gonna begin sourcing a few extra parts needed and I'll ask a few more questions about some details I am not clear on.
This
So as I examine the layout (now included), I am not understanding a few areas. The volume pot is not connected to the 10k down to pin 2 of the V3?
The Reverb pot connection on the input is not made?
On the switched cathode (5uf 2.7k) should that be going to ground and not up to the 68k string?
The schematic shows a connection from V3a pin2 to the reverb pot input? I am not clear on how that gets made in the layout.
Sure would appreciate any clarification on this before I proceed to making these changes.
-
Ok, so Tubenit was kind enough to address some of these questions on an older thread. Sorry if that's muddying the waters and might have been better to keep it all here.
I have an interim layout now for Gene's CBS schematic. I say interim as I now need to create a layout that can get inserted in the existing Bogen, albeit stripped back to some degree.
-
A question, I have 40uf for both B+1 and B+2 then down to 20uf, and 10uf down the line. I can decouple the caps to do 40, 20, 20, 10, 10 if ya'll think that's a better course.
It's playing in its current format (have not done the full conversion yet), ok.
-
IF...Your meter can handle 600VDC or VAC... measure the VAC at each PS tap
should be in the ballpark of 2-4VAC, <500mVac, <100mVac, <20mVac
-
I have an Extech 530 and when I try to read VAC at the PS (B+1) the meter dances back and forth between a reading that makes no sense? Can you go a bit more into that reading and what it represents? Leakage?
-
those "taps" are part of C-R-C filters, that "smooth" out the rectified DC volts, by measuring AC volts at the taps you can "see" how much AC is filtered out.
To do that you need a scope or meter that can handle the 400+ VDC AND be able to "discern" the AC volts "riding" on the VDC.
-
Ok looks like my meter cannot handle that. I have scope but it's been a while....
Here is the conversion layout. It's based on accurate dimensions of the chassis, tube sockets, Trannies etc. I would put the board on standoffs and include a backer board in key areas.
I would welcome any critique of this layout or suggestions.
-
PT HT ground preferably links to first filter cap neg lead, so that the charging current pulses do not flow through chassis.
Preferably use a single link from secondary 0V to chassis - such as shown next to input socket.
-
, B
PT HT ground preferably links to first filter cap neg lead, so that the charging current pulses do not flow through chassis.
Preferably use a single link from secondary 0V to chassis - such as shown next to input socket.
If I understand correctly, the CT of the PT should hit the same ground as the first filter cap. Then
The Ground of the secondary OT is it? Over to the input ground? Ok not that.
You mean a 'single link' on its own - got it.
:icon_biggrin:
On the Reverb circuit, I think I have it wrong....should it be Blue to V2 plate (pin 1)
Red to B+, Green to RCA input tip and Black wire to Sleeve of Input RCA?
Output RCA tip to Pin 7 of v2.
Isolate both jacks and ground the sleeves.
-
Updated layout
-
Here is an alternative layout. Less cross talk.
-
PT HT ground preferably links to first filter cap neg lead, so that the charging current pulses do not flow through chassis.
Preferably use a single link from secondary 0V to chassis - such as shown next to input socket.
https://www.valvewizard.co.uk/Grounding.html
-
PT HT ground preferably links to first filter cap neg lead, so that the charging current pulses do not flow through chassis.
Preferably use a single link from secondary 0V to chassis - such as shown next to input socket.
https://www.valvewizard.co.uk/Grounding.html (https://www.valvewizard.co.uk/Grounding.html)
I believe I have the first recommendation correct - the HT PT is grounded with the first Cap -
On the second suggestion, on my Marshall 100 Lead I took the speaker ground to the presence ground (Larry ground)
Here would it be better to ground the output over to the input ground?
Right now I have four nodes for ground
1. Input
2. Power rail and Power tube cathode
3. Mains ground
4. Output
Are you suggesting I eliminate the output ground?
-
You should read the grounding link I posted from Merlin.
I believe I have the first recommendation correct - the HT PT is grounded with the first Cap -
If #4 is the PT CT, then this is wired wrong. CT and 1st filter cap ground only go together, then a wire is run from there to the chassis ground or other grounds. You have everything grounded together like a star ground.
On the second suggestion, on my Marshall 100 Lead I took the speaker ground to the presence ground (Larry ground)
Here would it be better to ground the output over to the input ground?
No, the OT secondary ground lead has the most current in the amp. It goes with the power amp ground.
Are you suggesting I eliminate the output ground?
The OT secondary ground lead has the most current in the amp. It goes with the power amp ground. The presence pot ground goes with PI local star then is run with a wire/daisy chained back over to the power amp star.
I passed a kidney stone and I'm on pain meds, I can't do this until I feel better.
-
Ouch, sorry to hear about your health issue. Appreciate your taking time even with .....
I have read Merlin's grounding many times. I can say the quietest amp I have (most are) is the Trainwreck Rocket which is a single point ground. Then the Dlite is star ground and very quiet and many others that are variants. My 100w lead is star Ala Larry's. Let me try what you suggest.
Best A
-
With Merlin's grounding; each filter cap has it's own local star and those stars are then daisy chained to together.
You can then go with 2 chassis connections, power amp and preamp. Or you can do the single chassis ground.
The noisiest ground connection is the 1st filter/PT CT. And the speaker ground has a lot of current, so it can be a problem.
With a true star ground you end up with wires all over the place, very messy.
I have no idea who 'Larry' is.
-
So - Larry Novisiber - came up with a grounding scheme used widely, here I think and at Metro and TAG.
Thanks for the intel.
-
No thank you on Larry.
I would never do my grounds like that. Never.
And if you've read the Merlin grounding link several times, like you say, and understand it, you wouldn't use that grounding either.
-
No thank you on Larry.
I would never do my grounds like that. Never.
What I mean by this is, you can get a grounding layout to work on a certain amp, but that doesn't mean it will work for any/all amps.
Merlin's and KOC's will work with any amp. There are reasons they ground what with what and where.
-
Yes understood. I am mapping out my options. I know I am over-thinking this whole project, but for me it's a hobby and I am taking my time and want to make it great. I'll post a new upgrading grounding scheme soon. Work is now getting in the way... :icon_biggrin:
Thanks.