• 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 first post and a question

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
hendrixgr,

Do not use electrolytics for coupling caps. This is a tube amp. Think Meg Ohm grids, and think leaky electrolytics.

Lots of solid state amps use electrolytics. Think non Meg Ohm Transistor Bases. I know someone who went completely through a sound studio solid state sound board, and threw out all the electrolytic coupling caps, and replaced with film caps. Good call.

By the way, there are "bipolar" electrolytics, do not use those for tube amps either.

"All generalizations have exceptions"..
Thank you for the information! I have heard about those bipolar electrolytic capacitors but i have never used one.
You are correct about the leaky electrolytic capacitors especially as they get older and dryer, i didn't think of that!

Chris
 
I don't understand what you mean. The way i measured the ecc84 stage was by giving it signal from a HP Agilent audio generator and then measure the signal using a HP Agilent audio analyzer.

The signal was taken from the grid of the power tubes but without the output tubes installed if that is what you mean.
I only tested the preamplifier stage not the power stage yet as i need to order the output transformer made locally.
I have also connected a 220v to 6 volt transformer as a preamplifier output transformer at the grids of the output tubes (balanced signal fed to the transformer) and it gave a very good sound without distortion with enough power!

Chris
 
Last edited:
..........
the sine wave getting bigger and bigger as the tube is getting warmer but at some point after it reaches the 41 volt output the amplitude is getting smaller again and it settles at about 38Vpp but with 0,8% distortion now not 0,2 as the first valve...................

sir this ecc84 working in phase splitter.. power out put tube are like lode for tube in phase splitter.when sine wave getting big ...your out put tubes also getting hot and mean these tube like lode for output of ecc84.
so without output check .
another point: same tubes should be same in your oscop.
 
I only tested the preamplifier stage not the power stage yet as i need to order the output transformer made locally.
I have also connected a 220v to 6 volt transformer as a preamplifier output transformer at the grids of the output tubes (balanced signal fed to the transformer) and it gave a very good sound without distortion with enough power!
Chris

!!!!!!
you just asamble driver ?
but how you get good sound ? connect power trans as interstage transformer to output of phase splitter or first ecc84 ?
 
Ohh i see what you mean, i am testing without load! The preamplifier tube only draws 3ma per triode so it does not get very hot and the output tubes should not make a big difference as the grid resistors are already included.
Tomorrow i will check many tubes and report back.

Chris
 
Hi.
Good to know because the tube manual says that it is intended as an rf cascode amplifier but it does not say anything about remote cutoff.
Chris

They never do. Here is a loadline for the 6BQ7 I used as a cascode LTP. (Attached) See how asymmetric the constant voltage line (green) is: that means variable-u even though the spec sheet says nothing about this. The type was intended for use as a VHF cascode often appearing as premix RF amp for TV's and FM xcvrs. Despite that, 6BQ7s can work well as audio amps though finding good loadlines is a bit more difficult.
 

Attachments

  • 6BQ7A-Loadline.png
    6BQ7A-Loadline.png
    120 KB · Views: 164
If you do not have the output tubes installed, the B+ will almost certainly rise to a higher voltage than when the output tubes are installed. That means two things: The B+ filter caps might have their voltage ratings exceeded. The Driver stage will have more B+, and so may work differently (more voltage swing before clipping, different distortion, different power dissipation, etc.).
 
Hi. Each stage has it's own smps supply, voltages are regulated. BTW thank you for the variable mu explanation, i have a lot to learn as tubes are a stranger to me (but i am learning!) After some more calculations it looks like the sweet spot for ECC84 grid bias voltage is -5v @ Ia=3ma and 22Kohm anode and cathode resistors. Tomorrow i will see if the theory agrees with real measurements.

Chris
 
Last edited:
It looks like the first tube out of 10 that i used was the best and the second the worst of the bunch... All other 8 tubes were exactly the same just a little more distortion than the first with the average THD being around 0,3-0,4 % for 40Vpp output without any negative feedback, not bad i think.

I also changed the leds to high intensity white (blue will do also i think) which gave me a bias of 5,3v with two leds. 4 to 5 volts bias proved to be the best for the preamplifier and 5 for the phase splitter, the phase splitter's best load was 10Kohm (i tried different loadlines, bias and load resistors in an effort to check the valves datasheet). I have tried different versions with cathode resistors (always the same operating point and each valve stage separately) bypassed and not, direct biasing etc but the LED method proved to be the most consistent with less distortion not to mention the visual effect.

Chris
 
hendrixgr said:
BTW thank you for the variable mu explanation, i have a lot to learn as tubes are a stranger to me (but i am learning!)
Variable-mu valves are not as strange as you think. An ideal vari-mu has an exponential characteristic - just like a BJT! The difference is that the small 'thermal voltage' of the BJT is replaced by something much bigger, maybe a volt or two. However, no real vari-mu valve is as close to exponential as a BJT.
 
However, no real vari-mu valve is as close to exponential as a BJT.
That is correct sir! Today i ordered the output transformer (130Euros for 60w 6,6Kohm push pull with interleaved windings and 4-8 ohm output), my monoblock amplifier will be ready in two weeks.

I asked for a power supply transformer and they gave me a price of 80 Euros which is not bad for a 400w 500v@0,5A (toroidal or iron, i didn't ask) transformer with 250v@200ma windings (obviously i need either a bigger one or one for every channel.

Chris
 
Last edited:
short tail pair test

Today i gave a try to a ac coupled differential amplifier setup with ccs like the one found here Practical Phase Inverters. at the end of the page with one with the 2n3904 tail transistor and after some measurements i must say that i wasn't impressed.

The concertina with preamplifier works much better ( measurements taken with instruments not ear) and has a higher overall gain 15 vs 5 of the differential setup (about 1/4 of the mu) and has lower THD by 0,6% (0,8 instead of 0,2), not much but lower. One thing that makes me doubt my measurements was that the source of the added distortion was evident, one valve half had Ia= 5ma and the other half 3ma something not normal and not matter what i tried i couldn't make the two valve halves draw equal amounts of current.

Maybe the ECC84 does not contain two matched valves or the CCS i used was not that good for the LTP but fine as a plate load for the concertina preamplifier (i will change it eventually as i don't like this CCS topology at all) and yes it was powered with -20v so it had enough room to work. After all that the amplifier construction continues with the schematic topology i posted as i don't need the added output amplitude the LTP can provide.

Chris
 
Last edited:
I am not sure that ecc84 is suitable for an ltp ( or any use of 2 identical triodes) The g1 is also connected to the screen between the triodes for instance.

Balance between triodes in general may differ as much as you have observed, that is consistens amopng all models. Matching is reuired if you need more close match ( or use a circuit that is insensitive of this )
 
hendrixgr,

You can not plug in an ECC84 into a 12AX7 cathode coupled phase invertor circuit and expect it to work properly. Getting the filaments and rest of the pin-out correct is the easy part of the revised circuit.

You are starving the ECC84, it has much higher transconductance, much lower u (mu), and much lower plate resistance.

You must re-design the plate resistor (use much lower values), and the 2N3904 emitter resistor (use a much lower value). You need more current in the ECC84 than in the 12AX7. If you copy the circuit values on post #57, the ECC84 plates will be at a very low voltage.

As DF96 said, the triodes are built the same, but may not match. And grid 1 has much higher capacitance, it is in parallel with the shield (shielding triode 1 from triode 2). Grid 1 is supposed to be at AC ground which is why the shield is tied to it (grid 1 is the top tube in the cascode circuit). This will be OK for audio, but only if you drive grid 1 with a low to medium impedance; otherwise there may be attenuation at high frequencies for triode 1.

If you do not have ECC84s from several manufacturers, or at least several from the same manufacturer, you might not find any that have DC balance. Because the ECC84 has 6000uMhos transconductance, you may be able to use a 50 Ohm pot with the wiper going to the 2N3904 collector, and the pot ends to the cathodes (adjust the pot to balance the plate voltages). But that will only work if you stand quite a bit of current, otherwise there will not be enough drop in the 50 Ohms of the pot resistance (50 Ohms x 2mA is only 100mV; and 100mV x 6 mA/V = 0.6 mA. That is not enough adjustment to make up for the 2mA difference. Perhaps if you use low enough plate resistor values, and higher current in the plates, the tube match might be closer to start with, just because you are using a different part of the tube curves.

The problem is that if you get DC balance, you still have to get AC balance. You would have to use a large bypass cap across the ends of the 50 Ohm pot to "connect" the cathodes at AC. 6000uMhos is 166 Ohms, the cathode impedance is 166 Ohms + Rp/24. (the cathode impedance increases by the plate load divided by u). Just bypassing 50 Ohms @ 10Hz for -3dB requires 320uF, which is -1dB @ 20Hz. This capacitor can not be an electrolytic. If you use a bypass cap, then you can use a 100 Ohm pot, and get 2 times the DC balance adjustment range, and still have AC balance too.
 
Last edited:
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.