I guess we need to see the schematic of your amp. How different is it from the stock one posted above? 100mV on the cathode voltage is crazy. I see two full wave rectifiers- did you change that? You can see how the filament virtual reference is tied to the cathode voltage in the stock amp. Matching tubes is not done to minimize hum that I know of. Hum happens matched or not. Did to try shorting the amp input to ground to see if the hum is still there?
It helps with power supply ripple rejection as a matched set of tubes has the exact same gain, plate and screen current draw ETC... which means that any ripple will be in phase and the same voltage on both plates which means it gets cancelled in the output transformer.
Here's the schematic how it's currently wired.
Testing the amp with nothing plugged into the input I get no hum and no buzz.
Connecting the audio cable, but leaving it unplugged from the buffer I get a slight barely audible buzz.
The only noise I now get is a barely audible hiss (more of a rushing sound) similar to what one might hear with a solid state amp and it is only audible with my ear right by the speaker. That is acceptable to me given it cannot be easily heard.
I would absolutely love to get rid of the buffer after the preamp, but the only other way I know to get gain and a low impedance is with an OP-AMP which introduces solid state after the preamp and is something I don't want. Not sure why it really matters given my sources are solid state.
Connecting the audio cable, but leaving it unplugged from the buffer I get a slight barely audible buzz.
The only noise I now get is a barely audible hiss (more of a rushing sound) similar to what one might hear with a solid state amp and it is only audible with my ear right by the speaker. That is acceptable to me given it cannot be easily heard.
I would absolutely love to get rid of the buffer after the preamp, but the only other way I know to get gain and a low impedance is with an OP-AMP which introduces solid state after the preamp and is something I don't want. Not sure why it really matters given my sources are solid state.
Ok, so you finally answer my question- your sources are all solid state, but the pre-amp driving this amp is tube, (including the buffer) so you are stuck driving a long distance with high impedance output (47K). Correct? That matters, and is why I have been asking. And now you show a 47K to ground, which is also something that was talked about, and you did not say anything and only showed the 1 Meg.
Where did this re-design come from? Why did you reduce the gain of the input and phase splitter? Generally I would think within the feedback path you would want higher gain? I think a lot of this re-design is above my pay-grade, so pinch of salt filter engaged.
Well, anyway, I understand having a pre-amp with a unity gain if you have relatively hot sources, and the amp had more gain like it did stock. People usually have the opposite problem- hot sources into too high gain in the pre-amp, to a high gain amplifier. Your buffer looks like it has gain does it not?- the second stage 12AT7 is a follower, but don't you get any gain out of the first stage 12AT7?
I understand using a tube pre-amp if you had a shorter cable run. You have a long cable between a 47K driver and a 47K input load and that is going to be noise prone, but it has to be that way because neither your pre-amp or amplifier have enough gain. I think you have answered your question- I think you maybe need to build some modest gain into your pre-amp. How much depends upon what input signal amplitude your amplifier needs to put out the volume you desire. If you feel the volume levels are adequate when testing preamp+buffer and amplifier with a short cable, perhaps you need to decrease the output impedance of your buffer- perhaps parallel followers?
On the amplifier- The virtual center of the filament string should not be connected to the negative grid bias. Now I understand what you were meaning by that- I had no idea you would do all that work to create a separate negative supply for the grid when it is done so elegantly by elevating the cathode, I'll go with it. The filament center should connect to ground, since that is what all of your driver cathodes are so there is not potential there. By connecting this negative, you are maximizing that parasitic tube inside the cathode by making the cathodes positive compared to the filament heater wires.
With the driver grids grounded, I would assume the grid supply would need to be whisper quiet if you had non-matched tubes, and the two 100K pots were at different voltages (not being able to cancel out the ripple in the output by having the drivers respond exactly the same). The negative grid supply is easy as cake to make as clean as you want- You can pick any transformer to give you enough headroom, and regulate it with a linear regulator. Nobody will see you hid solid state inside. Or, use larger filter capacitors- these could be 1000uF @ 50V. I see it has a full-wave rectifier- is that true presently?
How do you adjust the bias current on this thing? You have no cathode resistor to measure the quiescent current through- Do you use the DCR of the output transformer or an ammeter? Also kind of curious about the screens being at a higher voltage than the plates?
My take away from this redesign is that in the end, it is basically trying to accomplish what Dave Gillespie does when he uses the Enhanced Fixed Bias, using a negative voltage regulator to hold the cathodes at a fixed bias across load. (I'm most familiar with the application in the Magnavox 9300 and it's wonderful)
BTW- You seem to have more than enough knowledge and skill to tackle this problem. I'm not sure how any of us could have really helped without seeing the whole picture, but it's been fun running around the obstacle course with you 😊
Where did this re-design come from? Why did you reduce the gain of the input and phase splitter? Generally I would think within the feedback path you would want higher gain? I think a lot of this re-design is above my pay-grade, so pinch of salt filter engaged.
Well, anyway, I understand having a pre-amp with a unity gain if you have relatively hot sources, and the amp had more gain like it did stock. People usually have the opposite problem- hot sources into too high gain in the pre-amp, to a high gain amplifier. Your buffer looks like it has gain does it not?- the second stage 12AT7 is a follower, but don't you get any gain out of the first stage 12AT7?
I understand using a tube pre-amp if you had a shorter cable run. You have a long cable between a 47K driver and a 47K input load and that is going to be noise prone, but it has to be that way because neither your pre-amp or amplifier have enough gain. I think you have answered your question- I think you maybe need to build some modest gain into your pre-amp. How much depends upon what input signal amplitude your amplifier needs to put out the volume you desire. If you feel the volume levels are adequate when testing preamp+buffer and amplifier with a short cable, perhaps you need to decrease the output impedance of your buffer- perhaps parallel followers?
On the amplifier- The virtual center of the filament string should not be connected to the negative grid bias. Now I understand what you were meaning by that- I had no idea you would do all that work to create a separate negative supply for the grid when it is done so elegantly by elevating the cathode, I'll go with it. The filament center should connect to ground, since that is what all of your driver cathodes are so there is not potential there. By connecting this negative, you are maximizing that parasitic tube inside the cathode by making the cathodes positive compared to the filament heater wires.
With the driver grids grounded, I would assume the grid supply would need to be whisper quiet if you had non-matched tubes, and the two 100K pots were at different voltages (not being able to cancel out the ripple in the output by having the drivers respond exactly the same). The negative grid supply is easy as cake to make as clean as you want- You can pick any transformer to give you enough headroom, and regulate it with a linear regulator. Nobody will see you hid solid state inside. Or, use larger filter capacitors- these could be 1000uF @ 50V. I see it has a full-wave rectifier- is that true presently?
How do you adjust the bias current on this thing? You have no cathode resistor to measure the quiescent current through- Do you use the DCR of the output transformer or an ammeter? Also kind of curious about the screens being at a higher voltage than the plates?
My take away from this redesign is that in the end, it is basically trying to accomplish what Dave Gillespie does when he uses the Enhanced Fixed Bias, using a negative voltage regulator to hold the cathodes at a fixed bias across load. (I'm most familiar with the application in the Magnavox 9300 and it's wonderful)
BTW- You seem to have more than enough knowledge and skill to tackle this problem. I'm not sure how any of us could have really helped without seeing the whole picture, but it's been fun running around the obstacle course with you 😊
The preamp uses a FET, but can switch it in and out of circuit so that the preamp can be completely passive.
It's not shown on the schematic, but I have 1 ohm resistors in each cathode circuit and each pair of tubes goes through its own analog milliammeter, however soon as the other digital milliammeter comes in I'll install both of those where the analog milliammeters were.
Yes the schematic is how it's currently wired.
The ripple voltage on the bias line was reduced from 100mV to 5mV by redoing the bias supply as shown on the schematic.
It causes a slight increase in plate voltage on the 6V6 tubes which causes a slight increase in output power plus it gets rid of that pesky cathode bypass cap and without that cap the bass is a little better.
From a Facebook group.
Also from the same group.
I do have knowledge and my knowledge is growing through forums like this one. I like to discuss stuff like this in case someone sees something I'm missing when I'm troubleshooting a problem and/or sees a problem I didn't know was a problem and/or knows a better way to do it than how I've done it.
Yes the buffer does have gain. The 50k trimpot adjusts the level.
The driver and phase splitter redesign came from a Facebook tube amp group and was done to solve the issue of the stock driver and phase splitter relying on both 12AX7 sections being closely matched and to fix the issue of the output increasing the closer I would get to 20kHz feeding a sinewave into the amp with the proper load on the output transformer.
The 5V4 was my choice as it has lower voltage drop than the 5U4 does which was needed to raise B+ slightly. The amp now does at least a solid 20 watts RMS.
Some of it is above my pay grade such as how the driver and phase splitter circuit works (I could look it up and learn how it operates) and figuring out why the output increased as I got closer to 20kHz. I'm not knowledgeable enough to know what the cause of that is and only know that I was able to fix it with a trimmer cap. It's something I've not seen on any other tube amp and I've built and rebuilt a few.
Last edited:
Best solved with a true low output impedance buffer with a quality opamp. Low part count, low energy use, end of hum. Drives 600 Ohm. No rational thing against it as sources are not tube devices anyway. It would be good to have the device with power on/off muting shorting outputs to GND for total silence.
The cause of the issue is of course the distance. That requires other measures than using a tube device to solve it adequately. The used tube is not the right one for that purpose either. Tubes are nice and all but really not the final word when putting high impedance stuff far from eachother. High impedance high gain stuff with non standardised line signal level should be stacked with short as possible cabling and only power amplifier output cabling as long wiring.
If you make it a stereo buffer you can connect a second Magnavox AMP-142 and enjoy stereo.
Last edited:
I could, however I only have room for one speaker in my bedroom.
An OP-AMP is definitely a good way to go, however I've got the hum and noise to where it's barely audible now so the only potential benefit is if there's any treble loss due to miller capacitance issues.
I likely should have used a 12AU7, but the 12AT7 was what I had and it works.
I'd love to modify the preamp to where it has gain, however I don't want to modify it given the preamp is good quality and wasn't cheap.
I'll definitely reconsider the OP-AMP idea as I know it will be better. What OP-AMP should I use?
Also I'll measure the input and output of the tube buffer so that I can calculate the gain.
The benefit to an OP-AMP circuit is i can use one OP-AMP per channel and have rhe right and left input totally isolated from each other and also that would enable me to make it stereo relatively easily if so desired.
An OP-AMP is definitely a good way to go, however I've got the hum and noise to where it's barely audible now so the only potential benefit is if there's any treble loss due to miller capacitance issues.
I likely should have used a 12AU7, but the 12AT7 was what I had and it works.
I'd love to modify the preamp to where it has gain, however I don't want to modify it given the preamp is good quality and wasn't cheap.
I'll definitely reconsider the OP-AMP idea as I know it will be better. What OP-AMP should I use?
Also I'll measure the input and output of the tube buffer so that I can calculate the gain.
The benefit to an OP-AMP circuit is i can use one OP-AMP per channel and have rhe right and left input totally isolated from each other and also that would enable me to make it stereo relatively easily if so desired.
I may indeed explore the OP-AMP idea further.
In keeping with the theme of potentially going stereo later I'll use a dual OP-AMP.
I'm thinking something like this will work, however I'm not sure what OP-AMP is best to use. I can just use a 9 Vac wall wart power supply.
In keeping with the theme of potentially going stereo later I'll use a dual OP-AMP.
I'm thinking something like this will work, however I'm not sure what OP-AMP is best to use. I can just use a 9 Vac wall wart power supply.
First: why do you need again gain? Sources are 2Vrms normally.
My guess is that the unknown preamp is an non standardised line level device. With 2Vrms sources one should only need source selection, a 10 …. 25 kOhm volume control and then an opamp buffer. More than enough output signal to drive a cable of 4.5 meter long to a tube power amplifier. Hum, hiss and miller issues will be non existing, trouble free operation the reward. The preamp seems a superfluous device electrically unable to do its desired task. Together with the either unable tube buffer that are 2 devices unfit for the desired application. All defined by real life demands.
This is a technical forum. Please avoid the “car for sale, color red” syndrome at all times. People want data as the OP both has the issue and the device. We drive blind!
My guess is that the unknown preamp is an non standardised line level device. With 2Vrms sources one should only need source selection, a 10 …. 25 kOhm volume control and then an opamp buffer. More than enough output signal to drive a cable of 4.5 meter long to a tube power amplifier. Hum, hiss and miller issues will be non existing, trouble free operation the reward. The preamp seems a superfluous device electrically unable to do its desired task. Together with the either unable tube buffer that are 2 devices unfit for the desired application. All defined by real life demands.
This is a technical forum. Please avoid the “car for sale, color red” syndrome at all times. People want data as the OP both has the issue and the device. We drive blind!
Last edited:
I'm afraid he has himself painted into a corner- he modified the Amp-142 power amp to have less gain in the input/phase splitter stage, amp has 47K input impedance, buffer has 47K output impedance, and a long run in between. Power amp now needs more input amplitude to achieve the desired volume. Tube pre-amp seems to have either unity gain or very low gain, and has high output impedance such that his solution was the 12AT7 follower buffer. I agree, poor choices of for desired application. The tube pre-amp is expensive and nice, the power amp is what it is (I don't think he has desire to tweak the front end), so I think a good clean op-amp buffer stage between the pre-amp and power amp is the least painful solution.
Tube Radio- Build the buffer with machine pin sockets, and don't crowd them, so you can test with cheap op-amps, and upgrade in the future.
As expected. Tube audio suffers most from not keeping to any standard by tube audio builders so that is the cause of most issues when interfacing with normal devices. All unnecessary but simply choices. Choices with consequences. And apparently always an extra tube device is thought to be a solution.
Anyway, I am confident suggested solution works way better. If I would need to make such choices to have my music playing on a mono amplifier/loudspeaker 4.5 meters away I would reconsider my options big time.
Anyway, I am confident suggested solution works way better. If I would need to make such choices to have my music playing on a mono amplifier/loudspeaker 4.5 meters away I would reconsider my options big time.
Last edited:
Most consumer grade solid state sources are under 1Vrms.
Thought I had mentioned the preamp earlier but I did not.
I'm using a Schiit Saga S preamp. The preamp is no longer sold, but here's the manual for it.
https://www.schiit.com/public/uploa...-iPmtf2s-ZCGTpa0W_DZkO4tboQB6axkrUrU-1w_dZoKd
Yes tweaking the front end of the amp is a no-go as I've got the amp to where it's basically flat to 20kHz and don't want to potentially upset things and have to spend time tweaking it again to be flat.
For the buffer it will definitely be built with sockets as I prefer that when building anything with chips.
The unmarked resistors will be selected once I figure out the gain of the tube buffer.
Thought I had mentioned the preamp earlier but I did not.
I'm using a Schiit Saga S preamp. The preamp is no longer sold, but here's the manual for it.
https://www.schiit.com/public/uploa...-iPmtf2s-ZCGTpa0W_DZkO4tboQB6axkrUrU-1w_dZoKd
Yes tweaking the front end of the amp is a no-go as I've got the amp to where it's basically flat to 20kHz and don't want to potentially upset things and have to spend time tweaking it again to be flat.
For the buffer it will definitely be built with sockets as I prefer that when building anything with chips.
The unmarked resistors will be selected once I figure out the gain of the tube buffer.
Last edited:
So it looks like the buffer was originally designed to be 1:10 impedance ratio, which is the common rule of thumb, for the original input Z of the Amp-142. That went over my head. So the 47K input is too low now. The buffer is supposed to be a ST to Mono combiner which in this case could just be done with a Y cable if the source device can give common line levels out. It seems the pre is used as a passsive device with no gain, but is it actually lossy? If the output Z of the buffer can be brought down to 10K, with enough drive, then the 47K could stay or maybe increased to 75K and still do the job.
Using a Y cable unless it has built in resistors doesn't work so well as both channels of the source try to drive each other.
Make it as complicated as desired by marrying dwarfs to giants but if it about good results you know what to do. Less is more.
Modern sources can put out 2Vrms so we calculate with that. Never issues with signal levels or impedances when keeping to standards. Only power amplifiers then need gain. Works quite good, saves a lot of time 🙂
What is the use of having a flat but low gain amplifier when the final results are mostly issues needing solutions (while still having a mono setup)!?
Modern sources can put out 2Vrms so we calculate with that. Never issues with signal levels or impedances when keeping to standards. Only power amplifiers then need gain. Works quite good, saves a lot of time 🙂
What is the use of having a flat but low gain amplifier when the final results are mostly issues needing solutions (while still having a mono setup)!?
Last edited:
Most all consumer audio sources I've dealt with cannot put out 2Vrms.
https://audiouniversityonline.com/consumer-vs-professional-audio-levels-what-is-the-difference/
So measuring the input and output of the buffer with just one channel connected to the buffer the input at 120Hz is 140.2mV and the output is 261mV which gives me a gain of 1.86 so I may make the buffer with a minimum gain of 1 and a maximum gain of 3 using a trimpot for each feedback resistor. That would require a Rin of 1k and a Rf trimpot of 2k. Maybe I can find a dual gang 2k pot and mount the control on the front so that both gains are adjusted equally.
https://audiouniversityonline.com/consumer-vs-professional-audio-levels-what-is-the-difference/
So measuring the input and output of the buffer with just one channel connected to the buffer the input at 120Hz is 140.2mV and the output is 261mV which gives me a gain of 1.86 so I may make the buffer with a minimum gain of 1 and a maximum gain of 3 using a trimpot for each feedback resistor. That would require a Rin of 1k and a Rf trimpot of 2k. Maybe I can find a dual gang 2k pot and mount the control on the front so that both gains are adjusted equally.
Last edited:
Not knowing how to deal with standards, impedances and gain structure is exactly why you have the issues. While the average medium is lower in level a test CD can cause 2Vrms. It would be foolish to design stuff so that it can not replay such media. Any modern DAC chip can put out 2Vrms. Since about 1982. Unfortunately recent DAC chips have way higher output level which is a pity.
Your power amplifier should be able to have volume control and then be directly connected to a modern low Zout source without troubles (or anything needed inbetween) not even with 4.5 meter cabling.
Your power amplifier should be able to have volume control and then be directly connected to a modern low Zout source without troubles (or anything needed inbetween) not even with 4.5 meter cabling.
Last edited:
In that case, as your buffer has 2 resistors for that, make a simple in/out Z adapter, maybe inside the 142 with one added jack.
What I posted is just the standard operating level.
I've personally never seen a typical consumer line level source go to 2Vrms.
Sure it's indeed possible, but none of my sources go that high.
That said before I build anything I may see what I have that could be made to work or what I can find on Amazon.
I've personally never seen a typical consumer line level source go to 2Vrms.
Sure it's indeed possible, but none of my sources go that high.
That said before I build anything I may see what I have that could be made to work or what I can find on Amazon.
OK. Check the specifications if only just out of curiosity.
Still make sure you change your setup to normal input sensitivity of the mono power amplifier as that is the root cause of everything you are experiencing. Trying to repair that with insertion of a extra gain stage after a preamp is a bit sad really.
Still make sure you change your setup to normal input sensitivity of the mono power amplifier as that is the root cause of everything you are experiencing. Trying to repair that with insertion of a extra gain stage after a preamp is a bit sad really.
Last edited: