Grandpa's Tube Amplifier / General Science and Electronics / Forums

Forums

4hv.org :: Forums :: General Science and Electronics

« Previous topic | Next topic »

Grandpa's Tube Amplifier

1 2 next

Move Thread

LAN_403

uzzors2k

Sun Jan 11 2009, 02:13PM

Registered Member #95 Joined: Thu Feb 09 2006, 04:57PM
Location: Norway
Posts: 1308

While visiting my grandparents I found some old tubes in a box which used to belong to my grandfather. Apparently they were part of an old "Radionette" radio, which is now long gone. I thought it would be pretty cool to recycle the tubes by using them in an audio amplifier. The tubes I found are 5Y3GT, 7C5, 7R7, ECH21 and what I assume is "EH2", but the name is almost worn off. The tubes seem ok, except the ECH21 which has a bent pin with cracked glass at the pin. The "EH22" has the same serial number on it from Radionette, so I'm assuming it was meant as a replacement. After some research I see the 5Y3GT is a fullwave rectifier and the 7C5 is a decent beam amplifier. I can't find much info on the other tubes, other than that they seem to contain a combo of either diode/pentode/triode based on various web searches. I think I can use the 5Y3GT and 7C5 without much difficulty, but I would like to use one of the other tubes as a preamp for the 7C5. This will be an amplifier for mp3 players or laptops. Now the question is which tube, and how do I set it up as a triode? The choice is between using the 7R7 or the "EH2" (ECH21 equiv I think).

Also, the tubes might be 60 years old for all I know. Is there anything I can/should do prior to using them besides washing?

EDIT: I forgot the links. 7C5, 7R7, ECH21 and EH2.

Proud Mary

Sun Jan 11 2009, 02:42PM

Registered Member #543 Joined: Tue Feb 20 2007, 04:26PM
Location: UK
Posts: 4992

You can use the pentode section of the 7R7 double-diode-pentode as your first AF amplifier to drive the 7C5.

Before using valves which have been out of service for many years, it is advisable to

1. Run the heaters only for several hours. Then:

2. Run them at reduced anode voltage and anode current for a few hours.

PS. You can use a zener to set the voltage on the screen grid of the 7C5, but you should still by-pass it with a capacitor.

Conundrum

Sun Jan 11 2009, 02:53PM

Registered Member #96 Joined: Thu Feb 09 2006, 05:37PM
Location: CI, Earth
Posts: 4062

cool.. how to use VFDs as valves. In fact it might be possible to run the individual segments as discrete valve elements for further amplification.

-A

Proud Mary

Sun Jan 11 2009, 04:53PM

Registered Member #543 Joined: Tue Feb 20 2007, 04:26PM
Location: UK
Posts: 4992

I think you are right to guess that some of the part number is missing from the valve EH2**.

ECH81 is a triode-heptode frequency changer, and the valve actually called "EH2" is also a heptode frequency changer (usually run as a self-oscillating mixer)

It seems to me very unlikely that a simple domestic receiver like your grandfather's would have had two frequency changers of any type (even if the receiver had VHF/UKW, a heptode would not be the valve to choose as self-oscillating mixer at this frequency)

So look inside your mystery "EH2??" valve with a strong light and see if you can see how many grids it has. I will guess that it is the RF pentode that preceeded the frequency changer, which would be the normal arrangement in a simple consumer radio like this.

I would guess the original line up had been something like this:

mystery valve (RF amplifier - almost certainly a pentode)

ECH81 (triode section: local oscillator, heptode section: mixer)

7R7 (D1: detector D2: AGC diode; pentode section: First AF)

7C5 audio output (equivalent to 6V6)

5Y3-GT as full-wave HT rectifier (originally working from a centre-tapped transformer)

Hope my guesses help!

uzzors2k

Tue Jan 13 2009, 06:49PM

Registered Member #95 Joined: Thu Feb 09 2006, 04:57PM
Location: Norway
Posts: 1308

The mystery valve is the 164884 one with the screen on the inside in the picture above. I'm not sure what to look for, but I'm sure your guess is correct. I've made a preliminary schematic, mind you this is the first time I've worked with tubes so there may be some grave mistakes. (Which I hope will be pointed out!) I based the design on the suggested schematics found along with tube info. I'm using the 7R7 as the preamp. I had to guess on some capacitor/resistor values, as I'm unsure of the function of many.

As for the output transformer. I need a 5W, 5k:8R transformer. I read a bit about making my own here, and given the low power demand I'm thinking of modifying a wall wart transformer. The power, voltage and lower frequency requirements are already met, and removing the secondary and winding my own won't be a problem. The primary inductance will be measured to ensure it is in range. Two things concern me however. 1) I'm running the output transformer in single ended mode, will it saturate? Do single ended output transformers need an air gap? 2) How does the DC resistance of the windings factor in? I see impedance for the xfrmrs is always given, but at what frequency is it measured? Is there a standard?

Proud Mary

Tue Jan 13 2009, 08:22PM

Registered Member #543 Joined: Tue Feb 20 2007, 04:26PM
Location: UK
Posts: 4992

Output Transformer:

If we divide the secondary load of a standard 8ohm speaker into the 5,000 ohm load resistance required by your 7C5 with 250V on the anode in Class A (see data sheet) we get an impedance ratio of 625:1. Now, if the impedance ratio is the square of the turns ratio, then the turns ratio is the square root of the impedance ratio. The square root of 625 gives you a 25:1 ratio to match your 7C5 into an 8 ohm speaker.

That was the easy part!

Much less simple, and usually requiring practical experiment, is trying to design and build a transformer having the flatest frequency response across the AF range.

Remember that a two valve amplifier is an oscillator just waiting to happen!

Don't forget to bypass the heaters to earth with, say, 1000pF, connected immedtately from the valve base to the chassis. Assume that there will be undesirable feedback (possibly up to the MHz range) and take measures against it. Decouple the HT line to earth with 1000pFs at regular intervals.

I will look at the data sheets for you, and check that you have set the screen voltage and grid bias correctly.

uzzors2k

Sun Jan 25 2009, 09:03PM

Registered Member #95 Joined: Thu Feb 09 2006, 04:57PM
Location: Norway
Posts: 1308

Thanks for the help, Harry. I've worked a little more at this over the weekend. Acquiring a center-tapped 450V transformer with a 5 and 6.3V filaments was going to be tricky, so I rewound a broken MOT I had. Putting a new ~200 turn winding on wasn't too troublesome. I just used a different MOT secondary as a 0.3mm wire source, and wound little 5 meter lengths of bifilar wire on pencils. These later joined after being wound on the core, giving a perfectly center-tapped winding.

I had two old Tandberg 4 ohm speakers kicking about, and I've decided to use the one as the chassis/box. It's slightly beat-up, smells moldy, and may not function anymore, but it looks retro. I've fitted the three tube sockets on, and the modified MOT. The little mains transformer is the proposed output transformer. The ratio is about right, and the 7C5 is only good for 5W so it should be large enough. As for linearity, we'll soon see.

One hitch so far. The 7R7 has a filament resistance of 80 ohms, is it done for?

Proud Mary

Mon Jan 26 2009, 01:23AM

Registered Member #543 Joined: Tue Feb 20 2007, 04:26PM
Location: UK
Posts: 4992

Heater resistance:

The hot* resistance of the heater should be near to 6.3/0.3 = 21R, so the cold resistance of a healthy heater should be much less than this.

By the way, in technical English, the word filament means the filament of a directly heated valve, while the word heater is used for the heating element inside the cathode of an indirectly heated valve such as yours.

*In the data sheets, some valve parameters are considered as being in either hot condition or cold condition

uzzors2k

Sat Feb 14 2009, 03:14PM

Registered Member #95 Joined: Thu Feb 09 2006, 04:57PM
Location: Norway
Posts: 1308

I purchased a new tube on ebay, which arrived earlier this week. It's a 7B6 triode which I bought due to availability, and the resemblance to the 7R7 in socketing and size. A while ago I was given a bunch of old fashioned resistors and capacitors from my aunt, which belonged to her father who was, or might still be a ham. Since they're so outdated I thought they would work well to complete the vintage feel. So after some work today the amp is done, and even better it actually works! Although not perfectly yet, I'm surprised just how well. Especially since the output transformer is just an old 5V cell phone charger transformer!

The problem is the speaker sputters and pops when the volume is set beyond a certain range. At low volume, and with the sound on my ipod turned right up it sounds fine. There's some background hum, but I assume some more filtering capacitors would take care of it. If the volume is turned up any more squealing and then popping noises come from the speaker. That's really the only problem so far. I would like a good amount of gain, since I want to use it as a guitar amplifier as well, though I'm not sure if that's feasible. Do you think the impedance and/or the gain is high enough? The 40 ohm resistors on the audio output are to load it down, since I've experienced some input sources like PC audio don't work without.

Here's how I've wired it up.

Aesthetics note: the volume potentiometer e will be fixed later. The blue switch has a small lamp in it, not a blue LED so it's still retro.

Proud Mary

Sat Feb 14 2009, 05:26PM

Registered Member #543 Joined: Tue Feb 20 2007, 04:26PM
Location: UK
Posts: 4992

Firstly, I see that you don't have any grid bias on V1, which is specified as -2V in the data sheet. At Ia = 0.9mA, you'll need 222R (220R 5% will do fine) in the cathode line to get your 2V voltage drop on the grid. To stop negative feedback reducing the gain, you'll need to by-pass the 220R with an electrolytic of 25uf or 50uF.

And then you need to put a grid leak resistance from the grid pin to Earth. (best, the same Earth point on the chassis as the cathode line.) Best results for the grid leak resistance will be somewhere between 100K and 500K, so I'd try 270K to start with, and see how that sounds.

Hum:

Your pi ripple filter capacitor sizes look as though you have just substituted the 680R resistor for the filter choke of a valve era design. If you reckon that V2 anode plus screen are drawing 54mA and V1 anode 1mA, then you are looking at a voltage drop of nearly 40V with those values of R and C.

I suggest replacing the rectifier valve with a solid state bridge rectifier (which will load your transformer less, because you won't need the rectifier valve heater winding) and then fitting an RC filter with values, of, say 100R and 1000uF, which will give you -35dB ripple attentuation but a voltage drop of only 5 volts.

Secondly, do you have the heater wires neatly and closely twisted as you should?

Have you routed the heater line well away from the grids of V1 and V2 as you should?

Is the high impedance input to the grid of V1 well screened?

Instability: The squealing you describe is oscillation, and the popping-type noises may be either the start up of oscillation, or could be due to poor supply line regulation, or both.

Are all leads short? Are outputs routed as far away from inputs as possible?

I suggest you also decouple the heaters to Earth with 0.047uF or so at the valve socket, which is always a good precaution against unwanted signals using the heater line as a highway around the chassis.

It wouldn't harm to put some extra filtration on the HT+ line - for example, 0.01uF and 1000pF in parallel with your electrolytics which can have significant impedance at higher frequencies. Also add some capacitors to Earth whereever the HT+ supply line looks a bit long.

I also suggest you connect the unused diode anodes of V1 directly to HT+ with 4M7 or so to keep them out of mischief - don't leave them 'floating.'

Try this out, Erik, starting with the HT+ supply filter and the biasing of V1 grid so that it is operating properly, as I've suggested, and you'll quickly start to see improvements!

PS: it's very normal in testing valve circuits to have to unsolder wires so you can insert your milliammeter and check that the current is what it says it should be in the data sheet. Choose only low impedance test points, so your meter leads and their capacitance don't become an active part of the circuit. You can also deduce the cathode current, for example, (which is the anode current plus the screen grid current) by measuring the voltage drop across the cathode resistor.

There can be considerable variation of parameters between even new valves - let alone those which are half a century old - so you must expect to have to do active voltage and current measurements when setting up a circuit to make sure that the valve is working under the optimum conditions defined in the data sheet. Don't expect any very exact correlation between the data sheet and what you actually measure in real life (except for valves designated as "Special Quality" ("SQ")and such) Everyday consumer valve circuits would use 20% resistors unless real accuracy was truly needed, which gives you some idea of the wide variation in operating conditions you can expect in real life. Good luck, lad!

1 2 next

Moderator(s): Chris Russell, Noelle, Alex, Tesladownunder, Dave Marshall, Dave Billington, Bjørn, Steve Conner, Wolfram, Kizmo, Mads Barnkob