Dedicated DIY Audio Thread

The Rocketry Forum

Help Support The Rocketry Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.
I just don't get it. An amplifier should just supply power to the signal without adding anything to the signal. Electronic amps are so close to perfect that you need to look elsewhere for the 'perfect' sound, especially the speakers.
If you *like* the distortion that tubes provide, there are ways to make power MOSFETs (in analog mode, not class D) sound like tubes. Much less expensive and far more power. If you are driving your amp into distortion, you need a bigger amp. IMO, your sound system should make the sound the same as the artist intended so look at speakers and the sound source.
 
Acoustics-wise, I'm mostly interested in noise-cancelling systems.
 
I just don't get it. An amplifier should just supply power to the signal without adding anything to the signal. Electronic amps are so close to perfect that you need to look elsewhere for the 'perfect' sound, especially the speakers.
If you *like* the distortion that tubes provide, there are ways to make power MOSFETs (in analog mode, not class D) sound like tubes. Much less expensive and far more power. If you are driving your amp into distortion, you need a bigger amp. IMO, your sound system should make the sound the same as the artist intended so look at speakers and the sound source.
Might be in the TLDR post category....

First of all thanks for watching this thread! I was right with you when I began my audio passion over 40 years ago and for most of those 40 years as an engineer.

From being a customer I got friendly with Paul Heath and Jim Gala (early highend dealers in Rochester) and spent alot of weekends and after hours at their places listening to many, many different components over the years. What struck me that components (amps, phono cartridges, CD players) all within the noise of each other in measured specifications all sounded different, some profoundly so.

In the case of amplifiers, some sounded glorious, some sounded sterile. Some provided spectacular 3D precision imaging, others provided a flat soundspace with no depth, some were just sloppy. Some amplifiers had me engrossed in the music, others left me just listening to the electronics. The dividing line was NOT tube versus solid state, there were excellent and poor experiences with both.

I was also a semi acquaintance of Roger Modjeski (actually he was just generous with his time on the phone with me and my buddy). In my opinion he was an engineer's engineer. He taught me it was actually easy to get ruler flat freq response and get to the limit of measurability of THD in any amplifier design. Just increase the open loop gain and provide as much negative feedback required to get the specs you want. But it will sound terrible.

Nelson Pass (Pass Labs) makes glorious solid-state sounding amplifiers that measure well and sound wonderful. His design philosophy is that it has to operate totally or mainly class A and minimize the NFB necessary by having a low distortion open-loop design. If I could afford or want to spend the $ on a Pass Labs amp this DIY thread would not exist.

As far a tube sound being due to added distortion, that may true in some cases but not in mine and that is an incorrect generalization. Below is the spice sim results of the amp I am designing for my next project. Its an UNSET push-pull using 1625 (WWII era) tubes. It will produce about 20W is class A which is more than enough for my 96db/1W speakers. In my last project the Spice sim was confirmed by measurements after completion.

This is the FFT predicting harmonic distortion. There is essentially zero 2nd order, and 3rd harmonic is over 70db down. So thats between 0.03 - 0.1% THD (The Mark Levenson No536 is rated at 0.3% THD). The AES has shown that these levels of distortion are inaudible in double blind tests.

So with the remaining years I have left and less with good hearing I will build a system that puts a broad smile on my face when listening to music (not the system). But I can also do it under a good measurement regime. So I can have my cake and eat it too. Plus tubes are cool.

1679859134318.png
 
Last edited:
Here is the schematic of the design referred to in the last post. One feature is the amp takes a balanced input (from my balanced outputs of my DAC) and the amplifier circuit is completely balanced from input to output. As opposed to a conventional design with converts a balanced input to single end internally. The balanced design eliminates ground loops, common noise on the interconnects and pretty much eliminates interconnect 'quality' from the equation.

Inputs are Vin+ and Vin-. Less than 1 volt drives the amp to full power. Power at clipping is about 20W in full class A. Also there is some 'sand' in the circuit (some tube purists hate that). There is a solid state current source for biasing of the input stage and the mosfet cathode drivers on the final output tubes.

1679929668082.png
 
Last edited:
This is the way.
Here is the schematic of the design referred to in the last post. One feature is the amp takes a balanced input (from my balanced outputs of my DAC) and the amplifier circuit is completely balanced from input to output. As opposed to a conventional design with converts a balanced input to single end internally. The balanced design eliminates ground loops, common noise on the interconnects and pretty much eliminates interconnect 'quality' from the equation.

Inputs are Vin+ and Vin-. 1 volt drives the amp to full power. Power at clippiing is about 20W full class A. Also there is some 'sand' in the circuit. There is a solid state current source for biasing of the input stage and the mosfet cathode drivers on the final output tubes.

View attachment 571180
 
My ears aren't what they used to be. Metal in the 70s and a manufacturing environment for the last 40 years have left me with hearing loss, particularly in the high frequencies. However, unless you do direct A-B comparisons with every thing *exactly* the same except the amp, I wouldn't trust the results. There is wide variation in quality of source and speakers. Don't even get close to clipping.
 
My ears aren't what they used to be. Metal in the 70s and a manufacturing environment for the last 40 years have left me with hearing loss, particularly in the high frequencies. However, unless you do direct A-B comparisons with every thing *exactly* the same except the amp, I wouldn't trust the results. There is wide variation in quality of source and speakers. Don't even get close to clipping.
That is what I did. I was shopping for a power amp. The speakers were one I already owned, Merlin VSBs. I auditioned Krells, Brystons, Music Reference, Thresholds, Aragon a bunch. They sounded decent, they all had 'perfect specs', they all sounded different. Would they generate identical results on 8 ohm power resistors? I'll bet they would.

Was it double blind? No. Could there have been some psychoacoustics going on? Maybe, I went with the Music Reference RM9 which I still have. It put a smile on my face in exchange for about $2500 at the time. I would have been happy with several of the others.

My 8 watt DIY 1625 amp produces a better listening experience for me than the RM9 with material that doesn't require bone crushing power. Acoustics, jazz and voices are just more realistic. When I want to rock out or play big orchestra I switch over to the RM9. When I want to impress my friends with 1812 Overture Cannon shots I will switch over to the Aragon solid state. When I play Tom Petty, Lou Reed or jazz quartets, the 8W 1625 is the one I hook up.
 
Last edited:
That is what I did. I was shopping for a power amp. The speakers were one I already owned, Merlin VSBs. I auditioned Krells, Brystons, Music Reference, Thresholds, Aragon a bunch.

My 8 watt DIY 1625 amp produces a better listening experience for me than the RM9 with material that doesn't require bone crushing power. Acoustics, jazz and voices are just more realistic. When I want to rock out or play big orchestra I switch over to the RM9. When I want to impress my friends with 1812 Overture Cannon shots I will switch over to the Aragon solid state. When I play Tom Petty, Lou Reed or jazz quartets, the 8W 1625 is the one I hook up.
good way to do it..bigger transients, extended range=more power
 
Right now I am in RM9 mode, Beethoven, Sammy Hagar, Kevin Mahogany Big Band, .. etc.
 
Here is a nice Tube and other DIY website. Pete Millett is a well regarded engineer all things audio focusing on tube but does dabble in solid state. If you want to dabble in a tube project for begineers, I recommend his "Engineers Amplifier". You can buy his PCB and assemble a VERY nice push pull tube amp for less that $500. His design is modifiable and adjustable to try various tubes and operating points. I will use it to rough out operating points and tube selection before a make a audiofool version. But it sounds excellent stock and has very low distortion.

His site.
http://pmillett.com/
Engineers Amp
http://pmillett.com/DCPP.htm
 
Finished utracer build and first tube curves measured!
IMG_20230406_183821127.jpgIMG_20230406_183829368.jpg
 
So now some tube matching results. When using more than 1 output tube in an amplifier its very important that the tubes have similar transfer functions, especially in a push-pull amplifier. In a push-pull amp you want the nonlinearities of the tube to cancel out leaving a linear response. This can only happen if the tubes are identical. So using my new utracer toy I measured basic tube characteristics for 15 of my about 17 1625 tubes. The values in the table below are:
Ia = plate current at 350 plate voltage and -20V grid bias
Gm = transconductance of the tube - di/dVg, ma/V
Plate resistance (hard to measure accurately in a Pentode
mu = amplification factor

I sorted the results by bias current. You see quite a bit of variability. The groups that are candidates for good matching are color coded. The groups that are in green far right are potentially a better matched group than the yellow. Next step to pick pairs out of these groups and run the full curves and find pairs where the curves fall on top of each other.

1681614634086.png

Bottom line is you cant get low distortion out of your tube amp by randomly selecting tubes. But if you do, the performance will be very good. Tube matching is not a myth.
 
So now some tube matching results. When using more than 1 output tube in an amplifier its very important that the tubes have similar transfer functions, especially in a push-pull amplifier. In a push-pull amp you want the nonlinearities of the tube to cancel out leaving a linear response. This can only happen if the tubes are identical. So using my new utracer toy I measured basic tube characteristics for 15 of my about 17 1625 tubes. The values in the table below are:
Ia = plate current at 350 plate voltage and -20V grid bias
Gm = transconductance of the tube - di/dVg, ma/V
Plate resistance (hard to measure accurately in a Pentode
mu = amplification factor

I sorted the results by bias current. You see quite a bit of variability. The groups that are candidates for good matching are color coded. The groups that are in green far right are potentially a better matched group than the yellow. Next step to pick pairs out of these groups and run the full curves and find pairs where the curves fall on top of each other.

View attachment 575201

Bottom line is you cant get low distortion out of your tube amp by randomly selecting tubes. But if you do, the performance will be very good. Tube matching is not a myth.
It's also why some audiophiles engage in tube-rolling, even if they don't realize that's why they're doing it.
 
It's also why some audiophiles engage in tube-rolling, even if they don't realize that's why they're doing it.
Or what they are doing. They put some new tubes, one is more mismatched, generates more 2nd harmonic distortion, sounds 'better' on acoustic, voices or small jazz. Then plays orchestral and sounds like cr*p (whats the 2nd harmonic of an orchestra?), then rinse and repeat. You can roll a tube but not the amount of negative feedback which has a different optimum for every tube type.
 
Or what they are doing. They put some new tubes, one is more mismatched, generates more 2nd harmonic distortion, sounds 'better' on acoustic, voices or small jazz. Then plays orchestral and sounds like cr*p (whats the 2nd harmonic of an orchestra?), then rinse and repeat. You can roll a tube but not the amount of negative feedback which has a different optimum for every tube type.
True, true.
 
It's not equipment, but it's DIY and audio-related-

Large 1970 Magnavox TV console, gutted for conversion to audio rack. Here's the backside:
_IMG_9122.jpg

And with a lower shelf fit:

_IMG_9126.jpg

The tall rectangular side openings held a 6x9 speaker in each side and the controls on the right side. I'm not sure what I'm going to do with those. The width is about right for CD's, so, I may make an LP stash on the bottom and a CD rack above that. It would hold a few of each.

Or put a preamp (active or maybe just a switch) in one side, if I can convince myself to be okay with the vertical orientation. It would also make a neat place to put mono amplifier power supplies, with L & R well separated. I think I've digressed, so far I've only cut and routed a piece of wood for it.
 
It's not equipment, but it's DIY and audio-related-

Large 1970 Magnavox TV console, gutted for conversion to audio rack. Here's the backside:
View attachment 575567

And with a lower shelf fit:

View attachment 575568

The tall rectangular side openings held a 6x9 speaker in each side and the controls on the right side. I'm not sure what I'm going to do with those. The width is about right for CD's, so, I may make an LP stash on the bottom and a CD rack above that. It would hold a few of each.

Or put a preamp (active or maybe just a switch) in one side, if I can convince myself to be okay with the vertical orientation. It would also make a neat place to put mono amplifier power supplies, with L & R well separated. I think I've digressed, so far I've only cut and routed a piece of wood for it.
I hope you pulled and saved the tubes when you gutted that tv. The xformers too. :)
 
I hope you pulled and saved the tubes when you gutted that tv. The xformers too. :)

I'm betting he didn't. :eek:

I’m a pack rat, I even saved the 6x9’s. They sound pretty good with vocal-heavy music, I like them better for that purpose than the 2-way KLH’s that are in the shop. And I saved the trim bezel that went around the TV tube, it’s hanging on my wall.

It was tough to make myself neck that big tube and gut it. The thing was in great cosmetic condition, bought from the original owner along with a Magnavox Stereo Theater console. But I just didn’t see myself bringing a big old TV back to life.
 
It was tough to make myself neck that big tube and gut it. The thing was in great cosmetic condition, bought from the original owner along with a Magnavox Stereo Theater console. But I just didn’t see myself bringing a big old TV back to life.
The sweep tubes (hor and vert) can bring decent nice dinner money on ebay if they are the right type.
 
The sweep tubes (hor and vert) can bring decent nice dinner money on ebay if they are the right type.
Sweep tubes make decent audio power amplifiers. They are designed to be very linear as they drive the horizontal and vertical raster. If the circular test patterns on the screen were round then the amplifier tubes were nice and linear......

$2 dollar sweep tube = $100 KT88
 
Sweep tubes make decent audio power amplifiers. They are designed to be very linear as they drive the horizontal and vertical raster. If the circular test patterns on the screen were round then the amplifier tubes were nice and linear......

$2 dollar sweep tube = $100 KT88
Heck, sweep tubes have even been used as linear amps for ham radio, in some cases.
 
Heck, sweep tubes have even been used as linear amps for ham radio, in some cases.
Yup, and illegal CB linears help dry up the supply of good ones. They were $1 each and the home-brew designs ran them at too high a plate current and burned up the supply, why not at a buck a tube, until they were all gone.
 
Yup, and illegal CB linears help dry up the supply of good ones. They were $1 each and the home-brew designs ran them at too high a plate current and burned up the supply, why not at a buck a tube, until they were all gone.
I certainly hope you don't think I was referring to a CB linear.
 
I certainly hope you don't think I was referring to a CB linear.
No, I recognize the Ham amplifiers. They were the few that actually respected the tube ratings. But near the end of the TV tube era tube manufacturers were sticking various plates in the sweep tubes. The same tube, like the 6GB6GA, has 18W plates, some manufacturers put 7027 35W guts in them, same as the 6L6GC. So some linears were run above their ratings with no problems, others red plated and died because they were different tubes but had the same part numbers. Strange times.

http://www.vacuumtubes.com/6BG6.html
 
Back
Top