• WARNING: Tube/Valve amplifiers use potentially LETHAL HIGH VOLTAGES.
    Building, troubleshooting and testing of these amplifiers should only be
    performed by someone who is thoroughly familiar with
    the safety precautions around high voltages.

My Father’s Amplifier

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
In the late 1960's, my father wanted a nice stereo. I was in High School, finances were tight, he was working two jobs. He wished that I would figure out how to assemble a modestly adequate amplifier from odds and ends. I read Crowhurst’s articles from the 1950’s.

From time to time, I worked intensively on the project. Alternating with periods of intense study, my project lay fallow for years.

My father passed away in 1980, but the goal endures. With a well-aged B.S. in Physics and 47 years, I hope I have put it together. Physics or not, electronics is still sometimes opaque to me.

Several months ago, I retired. I now have time.

“The management question, therefore, is not whether to build a pilot system and throw it away. You will do that. […] Hence plan to throw one away; you will, anyhow.” Frederick Brooks The Mythical Man-Month (1974)

I am designing this thing myself. That makes it an experiment. As it is experimental, I am open to changing every part of it, given persuasive arguments.

Total amplifier power must drive Cerwin-Vega speakers hard enough to fill a modest listening room – 15 feet by 15 feet. Anything less than three or four watts per side is inadequate. If circumstances limit maximum actual power to less than 10 watts, then so be it.

Enough blathering. I have made some preliminary choices, after much chasing of rabbits down rabbit holes. I see carrots in my nightmares.

A transistor op amp will drive a 6N7 splitter, which in turn will drive a 6CD6 Push-Pull output stage.

6N7 is an older dual triode with internally connected cathodes, gain of 35.

6CD6 is a horizontal TV sweep tube.

I do not plan to use global negative feedback. I do plan to use local feedback of various kinds in all of the stages.

This whole thing has to start somewhere, so I start here, at the end: (I know this is wrong, but now how much it is wrong, or where)

20 Watt amp with two mono-blocks

Output stage:
6CD6 output pentodes
circuit configuration –
typical B+ feeding Push-Pull through center-tapped output transformers to tube plates, fixed bias grid, with shunt feedback
400 Volt B+ supply, regulated
screen @175 Volts,
grid bias @ -35 Volt
quiescent current 100mV

Splitter stage
6N7 Dual triode - gm of 35, as opposed to 20 for 6 S N 7
circuit configuration:
differential amplifier with a constant current sink tail
300 Volt B+ supply
grid bias @ -6 Volt
Quiescent tail current 10mV
gain to be determined

Front end
circuit configuration:
input transformer to both sides of a differential amplifier.

I am starting this thread in the Solid State forum, and will start a new thread in the Tubes/Valves Forum when the time comes.

The first step is to sort out the preamp.

The front-end will be a transistor op amp assembled from discrete components, with Bipolar Junction Transistors, (BJT) and plain diodes as active devices.

I first read Self’s paper on the blameless amplifier in 2001. Since then, I have been planning on using a variation of Douglas Self’s blameless amplifier front end – a differential pair with a Wilson Current mirror and no Voltage Amplification Section (VAS).

Since then, in the last ten days, I have found Samuel Groner’s preamps. Groner’s scematics have the right “feel”, the feel of basic competence.

Analog Circuit Design · Samuel Groner · Resources · Discrete OpAmps

What is the general reputation of Samuel Groner’s preamp circuits, e.g. the SGA-SOA-2 Simple Discrete Operation Amplifier version 2?

I consider 7 transistors, 8 diodes, 5 capacitors, 11 resistors and a choke on the high side of simple. 14 transistors plus necessary other bits is complicated, in my opinion.

Groner’s preamp circuits are somewhat more complicated, but probably address shortfalls I didn’t realize.

Or is this the wrong approach entirely?
 
It depends on your sonic goals. The entire concept will not work for me better than a Sony radio, but we are all different.

Startlingly, you go into great detail about trifles and sonic destructors and make no mention of the iron. Which, in tube projects is central.

I ask for advice and guidance. I am two giant steps removed from a rank newbie. There is much I do not know, and a little that I do. I am also certain that some of what I "know" is wrong.

If this thing performs no better than an old Sony radio, I will be disappointed.

As for the "iron", I make do with a limited budget. If I have to start over with output transformers, I will. In that event, it may be better to ask a moderator to move this thread to the Tubes and Valves forum, or even start a new thread there. I will do that, if necessary.

I thought a description of my motivations and goals would be helpful to understand my point of view. If I bored you, forgive me, it was not my intent. Please feel free to ignore that part.

Please be so kind as to briefly point out which parts of my approach are trivial, which are sonic destructors, and why I should not be asking about a solid-state front end.
 
If this thing performs no better than an old Sony radio

I can't speak for the DIY opamp front end. That is not my area of expertise.

It is possible to make a first class tube back end from the few bits you mentioned. It is also possible to make a mess with the same set of parts. Some details.

Driver stage:

6N7 Dual triode - gm of 35, as opposed to 20 for 6 S N 7

35 VS 20 is the Mu, not the Gm. The Mu is basically the theoretical maximum gain possible from a given tube when driving an infinite load (impossible).

differential amplifier with a constant current sink tail.....
gain to be determined

Best case gain in a diff amp will be roughly 1/2 to 3/4 Mu if both sides are driven, about 1/4 to 1/3 Mu if one side is driven. Gain WILL be reduced when local feedback is applied.

Quiescent tail current 10mV

Current is measured in milliamps. 10 mA would be about right for a 6SN7 or a 6N7.

Output stage:

400 Volt B+ supply, regulated

Don't regulate the B+ in a pentode amp. Regulate the screen voltage supply and the bias supply. These control the current through the tube.

screen @175 Volts,
grid bias @ -35 Volt
quiescent current 100mV

With 175 volts on the screen and -35 on the grid, the tube IDLE current will be about 50 mA per tube. I assume that's where the 100mV comes from. 50 mA is a bit too much for a 6CD6. The bias should be adjustable (say -30 to -50 volts) to account for tube differences. Then I would set it for somewhere around 30 to 40 mA. Choose the lowest current that makes good sound for long tube life. The peak currents will be far higher than this.

You didn't mention output power requirements except that 3 or 4 watts would be adequate, and 10 watts acceptable.......a pair of 6CD6's on 400 volts will make anywhere from 25 watts to 70 watts from 400 volts depending primarily on the choice of output transformer.

Figure out how much power you want, cost of the transformers are the big factor here. A set of 25 watt transformers and the power transformer to feed them is far cheaper than a set of 100 watt OPT's and a really big power transformer. The power level will determine the required transformer impedance. The physical size of the OPT's will determine the cost and bass performance.

For a pair of 6CD6's on 400 volts my guesstimator says:

25 Watts Per Channel will require a transformer in the 8000 ohm range.
35 WPC wants about 6600 ohms.
50 WPC wants about 5000 ohms.
60 WPC wants about 4000 ohms
70 WPC wants a 3300 ohm OPT.

I am currently working on a dual differential pair driver circuit and doing some testing with a bunch of different tubes. There is a box of 6CD6's waiting to be tested. I can't say when I will get to them, it could be a few weeks or months. When I do, I will have exact power, bias and drive requirements for them. My driver tubes are 6CG7's which are 6SN7's stuffed into a smaller bottle. The details are here:

Tubelab Universal Driver Board, 2015 version
 
To Moderators

It seems that this thread should be in Tube/Valves. If you agree, please move this thread to that forum.

To George at Tubelab

Thank you for your concise observations. I avidly follow your posts, I cannot say how many years it has been, but well before I registered on this site.

I will turn my attention to finding a pair of 8000 Ohm output transformers, and an appropriate power transformer.
 
Well best of luck.
I still have a tube amp, and 4 tube driven organs, but the tube power era is over. Out of the matched quad of output tubes I bought in 2011, one shorted after 20 minutes. That was from the factory in the country that is benign. I'm upgrading the ST70 to improve the sound (good for 1961, worse than two of my transistor amps) but fighting obsolescence of the 7199 input tube , and hoping the new driver tubes I'm buying for the new driver board stay in production for a few years.
I've never bought any defective transistors, having stayed with newark(farnell) mouser digikey. So the era of great bj transistors is ending, but still going on.
If you haven't bought all these parts, I'd try to start tubes with a tried & true kit, instead of designing everything. New design is for amp two or three. I recommend the transformers from triodeelectronics.com for their ST70 kit, although I like the triple triode driver (6SN7) board from classicvalve instead of triode's EF86 driver board, which has a single sourced pentode. Tubelab's driver boards may be fine, he didn't come up in my discussions previously. For your output, 6CD6 is a minority tube, which is not being supported by the guitar amp people that kept the tube factories open. 6CA7/EL34 6L6, EL84, KT88, that is what people are selling a lot of. Output tubes have to be replaced every 8-10000 hours in my experience, rectifier tubes every 5000 hours. I go through 2000+ hours a year on my main amp. The cool thing about the ST70 kit, dynaco told you how to place every wire to avoid hum. The manual is on line at dynaco doctor. That is not trivial, looking at how hammond clock built their organ amps, they were very very careful in their construction plans.
Transistors can sound fine, I'm very happy with the sound of my Peavey CS800s although it needs repair for a channel that drops out (probably geriatric e-capacitors or relay contacts). If you want to get your soldering problems out of the way before you fight your design problems, I suggest you buy a honeybadger 200 w/ch transistor kit or something . I find 1 watt fine for base level listening, but as a classical listener I do have peaks of 120 W (26vac) in my material, and that is using 101 db@1W1m speakers, which are super loud.
The main problem with direct connected transistor amps is that one bad solder joint can blow a speaker driver with DC voltage. So build a protection board, one that disconnects the speakers with fets or the RJKeene way, which disconnects the output transistors from the power supply when the first one shorts out, to save the other three or 19 from meltdown.
Or build a capacitor coupled amp, one of which I have super-modified to sound much better than my tube amp.
Or buy a blown PA amp with serious protection like models from Peavey, Crown, QSC, Yamaha, and repair it back into shape.
BTW tube amps can melt down a tweeter with oscillation so make sure you don't have any. I blew two tweeters with the ST70, never detected ultrasonic oscillation but I don't really know that it didn't oscillate because my detection method was not instrumented.
As for the preamp, building up 14 transistors is a lot of work when IC's can sound really really good. Keeping the hum & radio interference out if challenging enough even with a stock circuit. My RA-88a disco mixer sounds better than my vacuum tube PAS2 after I super modified the RA-88a with $.38 op amp IC's, and pulled the hum inducing power transformer, fuse, and switch, out of the case. Also the 12AX7 PAS2 had some paper caps fail, which are also made of unobtainium and the extreme hint, don't replace .2 uf paper caps with .22 uf polyester, it sounds bad. Once again this is the polyprophylene cap era, so listen to the experts.
Having vacuum tube, transistor, and IC amps and preamps that can drive the same speakers is kind of a fun experience. I've got two listening rooms in my house, music and TV rooms, and a summer camp with another listening opportunity currently being satisfied by a $6 computer speaker, which obviously can be improved on.
So do what you want, but I have got two operating systems without going crazy, buy upgrading what junk I could buy rather than designing everything from scratch. I listen to music >3000 hours per year, so listening is more fun than designing is my hobby. I'm designing, but while I wrestle with the problems, I have systems to listen to.
Best of luck with your hobby.
 
Last edited:
I assume that you are looking to build a tube power amp because that is where your mindset
was in the 1960s? Because, as you probably know, it is cheaper to build a SS amp.
Would you consider switching to a SS power amp?
If it must be tube the simplest thing is to go to tag sales or check local classifieds to find
some bargains even if all you do is strip them for parts in order to build your own design.
First check if they have any collectable value of course.

You should probably look at the 100 buck guitar amp challenge and you'll find many tricks for
building a cheap tube amp. Output tubes for example, 6V6's are expensive because they
are used in guitar amps, 12V6 and 5V6 are not expensive. There are even EL84 types
with different heater voltages and pin outs that are cheap. I'd use these rather than
trying to use sweep tubes. On the other hand you might do just fine with Tubelab's guidance
and sweep tubes.
There are Dynaco replacement output transformers that should work for your design.
 

PRR

Member
Joined 2003
Paid Member
“The management question, therefore, is not whether to build a pilot system and throw it away. You will do that. […] Hence plan to throw one away; you will, anyhow.” Frederick Brooks The Mythical Man-Month (1974)

Even better than building a pilot system to learn from then throw away---

Study someone else's pilot system.

Brooks was working near the leading edge of his technology. Nobody was ahead of IBM. Your goal is very much in line with 70 years of audio systems. There's thousands of "pilot systems" to plagiarizXXX study.

Nobody can tell you what to build, any more than what lover to take or what car to buy. My neighbor sneers at my Honda and I don't see why he bought a Ford. That's your preferences.
 
I will turn my attention to finding a pair of 8000 Ohm output transformers, and an appropriate power transformer.

There are the reproduction Z565 transformers used in the Dyna ST35/SCA35 that should work for you. 8K. Ultralinear tapped for EL84s. Z-565 OUTPUT TRANSFORMER

I'd advise building something sooner than later that works - and then iterate on it. You can save the perfect chassis and perfect sockets and binding posts etc for later once you've settled on something. I started with plywood - and my HV supply (for 845s) was built inside of a milk crate. It was much later that I moved to custom CNC top plate and walnut box... I would even start with a Dynaco ST35 clone (cheap parts) and iterate on it - cathode bias to regulated fixed bias. Ultralinear to triode with regulated screens, etc.

-- Jim
 
Pete, PRR, and JRD:

Thank you. Your posts are noted and appreciated. I will look into your suggestions.

Sometimes it takes a while for me to post a response. It is not a lack of civility.

I intend my posts, including replies, to be informed and considered. That can take time. With a little luck, some may even be correct.

Please be patient with me.

Andy
 
Here's the description, note that good quality 7199s have gotten very expensive, but you
can rewire the socket for low cost TV tubes such as the 6u8 and a few others. Rewire the
outputs for low cost EL84 replacements:

"Dyna SCA-35 tested, needs resto $100

I tested it and made a single adapter for 7199 to 6U8 (I recall) to test one channel at a time
Has a replacement power transformer and some funky chassis mods, but its all there.
Wonderful transformers - Z-565"
 
Pete:

My Cerwin-Vega's are not the greatest, but they will have to do for now. When I want more power, I have a Yamaha AX-400 amp. Nominal advertised power is 50 Watts. I don't believe that for a minute, but the speakers distort badly long before the amp reaches half-full honk. For now, it is what it is.

I took JRMedford's suggestion and ordered from Dynaco. The other amp source you suggested sold the amp before I reached him. Thanks anyway.

I am working on my electronics bench right now. I don't fear 400 Volts, but I sure as heck intend to respect it. This isn't 9 Volt transistor work.

Isolation transformer, variacs, additional light fixture, oscilloscopes, independent adjustable power supply, signal generator. Building a wooden subfloor to insulate from the concrete basement floor.

I still need to round up lumber for the bench itself, a wooden chair, and a face mask. Got lots of goggles, but they tend to steam up my glasses.
 
I have no problem mixing sand and tubes. currently I see no better power tube then a superTriode config consisting of a 6922 triode and transistor setup. The 6922 curves are very linear. it is the driver for the tone stack in my altered Pioneer sx 780. the system requires 15 v p-p to reach full power output of 70w. I replaced the amp and adjusted internal regulator to get 70w into 8 ohms from the 780.
 
He wished that I would figure out how to assemble a modestly adequate amplifier from odds and ends.
That is a pretty low hurdle so this should be easy.
My advice is keep it simple. Figure it out for yourself - and listen! Learn from designs that have a proven track record for playing music convincingly.
Or is this the wrong approach entirely?
I tend to agree with analog_sa if you were trying to build an audiophile quality amp.; your information sources are poor. But you want to design an "adequate amplifier" so it doesn't matter...

...but if you were seeking audiophile quality ;)...valves are excellent voltage amplifiers, transistors are better current amplifiers. You propose valves in your current stage and transistors in your voltage stage, and a lot of them.
Valves can be excellent in the output stage, but you have a high cost and effort in the transformer. Transistors are a lot more difficult to make music with than valves; much cheaper but you have to apply a lot more effort.
Perhaps the best combination of the two technologies that I have heard is the Lamm M1.2
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.