modify parallel push-pull EL84 to single push-pull

[chungtat]

Recently I bought a integrated amp. The power output stage is using parallel push-pull EL84. Total 8 tubes.

The output transformer is 4K pri. impedance.

I don't need that much power and considering using only 4 tubes. That is changing from parallel push-pull to single push-pull.

What modifications do I need to do without changing the output transformer? Bigger cathode resistor? Or no need to modify anything?

The B+ is kind of high, 450 VDC.

Thanks for any help.

 

[ray_moth]

You will need a higher load impedance. I doubt if you can get away without changing the OP tranny, unless you can use different secondary taps for the speakers, e.g. connect 8 ohm speakers to the 4 ohm tap.

 

[chungtat]

I have forgotten to mention that the output transformer is ultralinear.

Tks.

 

[chungtat]

It does has 4 ohm taps.

One thing I dont understand is that even the ratio is correct, the load impedance still doesn't match.

 

[ray_moth]

. . . even the ratio is correct, the load impedance still doesn't match

How do you know it doesn't match? The normal p-p load I've seen for a pair of EL84 in UL PP is 8k. Connecting 8 ohm spkr to 4 ohm tap on a 4k p-p OPT should give you 8k. Just be careful if the paralleled EL84s share a cathode resistor. You will need to double the value to get the same bias voltage, since you will be halving the current.

 

[rdf]

If you don't need that much power why not try strapping them triode? Assuming the driver cleanly handles the increased grid capacitance.

 

[chungtat]

Thanks a lot! I will try that.

It is using parallel PP and I thought that make it single PP may have a positive effect on the sonic quality.

The cathode resistor is 10R for each tube. So I will double it for the 4 tubes.

What is good for cathode resistor for output stage? WW or carbon composite?

Thanks for your help.

 

[ray_moth]

The cathode resistor is 10R for each tube. So I will double it for the 4 tubes.

Hold on! I don't follow your logic. If there is really a separate 10R resistor at each EL84 cathode, then there is no need to double anything. Also, with only 10R present, it suggests that the amp is using grid bias, not cathode bias, which would probably require no adjustment to change from parallel to single PP.

Also, I don't believe changing from parallell to single PP will make any audible improvement to the sonics. Indeed, the opposite is more likely to be true. If you want better sonics, then it might be better to try rdf's suggestion.

 

[chungtat]

Here is the schematic of the output.

 

[analog_sa]

Also, I don't believe changing from parallell to single PP will make any audible improvement to the sonics.

It very probably will. I have never heard parallel tubes sound better. For reasons i don't understand parallel EL84 sound particularly bad.

 

[ingab]

Can you reseed the attachment (is 0 byte!!!) or e-mail at my address:

andrea.bonelli3@tin.it

Thanx in advance

Andrea

 

[chungtat]

Here is the schematic again.

It is grid-bias with a 249K resistor connected to a -23V DC. This part is out of this page.

I am thinking of turn it into cathode bias. But the B+ in this schematic is 450V and I can't find any reference with B+ of 450V.

 

[Johan Potgieter]

Just a mite from me:

For some reason I could not download your circuit. I have tried to locate EL84 data that I have here somewhere, but also unsuccessful. But as I recall working an EL84 into 4K instead of 8K lowers the max. output to about 65% of the 8K value, with some rise in distortion. What mainly worries me is the quite high 450V on your EL84s. Not that they will spark, but at a maximum dissipation of 15W (anode + screen), you can draw a max. current of 34 mA/tube. Is this happening?