• 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.

El84 Pp

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Hi everybody!
what do you think of this Dynaco lookalike amplifier with original SCA-35 OPT:s?
Should i use another input stage?
Can i easily convert it to be driven by E88CC/6922?
 

Attachments

  • dynaclonex.jpg
    dynaclonex.jpg
    62.9 KB · Views: 629
My 2 cents worth (well, depending what currency you talk of!):

I would watch it with the 12AX7 as an input tube. High mu and Miller effect (a-g capacitance) could play havoc with feedback stability/high frequencies, depending on the generator (pre-amp) impedance. This is especially so if your input comes from a pot (volume control) where this impedance goes from zero to a high value. In that sense I would prefer the E88CC or other lower mu triode, if you can provide the higher input. (My favourite circuit is an input pentode followed - directly connected - by a Schmitt phase inverter, although that means an extra tube.)

Comments?

Johan
 

PRR

Member
Joined 2003
Paid Member
> 0.24x300 comes out to 72W. Driver might take another couple of milliamps. 100VA ought to do OK

Disagree.

Transformer current rating is for resistive AC load, not a capacitor-input rectifier load.

But in trying to back myself up, I found different formulas. The RMS is at least 1.8 times the DC. But it can be higher. Generally in hi-fi, it is much higher.

For cores below 100VA, the real problem may be poor regulation, exagerated by capacitor-input spikes, leading to a much lower output than you expect.

For cores above 100VA, heat in the windings tends to be the limit. Heat is much worse than you would guess from the DC current. The very spiked current waveform has an RMS value far higher than the DC current delivered. Using rough estimates of winding and other parameters, 0.24A DC implies peaks over 1 Amp. The RMS value computeed by SPICE is about 0.640A in the winding for 0.230A in the DC load, a 2.8:1 ratio of RMS/DC, with a 100VA core with 10% rated regulation, 100uFd cap. And for 220VAC rated transformer voltage, B+ is 285V, somewhat less than Triactic's quick-guess said.

So even if 285V is acceptable, we need 220VAC 0.64AmpsAC, or 140VA. FWIW, my long-time rule of thumb is that core VA should be twice the DC Watts, and that happens to be exact in this case, except the output voltage is 285/325= 0.87% of what we hoped for. Correcting for that, we need more like 182VA transformer rating.

Double the DC Watts and round-up generously.

In sand-state and with some fixed-bias AB tubes, the idle current is low and the transformer heat is not maxed-out in speech/music work. But this self-bias affair will run near full current all the time.

RMS current and winding heat is MUCH lower with choke input. But a choke is a second transformer, just wound and gapped different. And since transformers are mass-commodity parts while chokes have fallen out of fashion, a more-VA transformer is generally cheaper than a tranny and a choke.
 

Attachments

  • rect-rms.gif
    rect-rms.gif
    8.2 KB · Views: 399
Take a good look at Eric Kingsbury's PP 6V6 amp. There is a LOT of food for thought in that design. If loop NFB is desired for UL or pentode operation, the grid of the splitter's non-inverting triode can be lifted from ground and the NFB applied to it. If that grid is lifted, the load resistors on both 1/2s of the 5965 should be the same value.

Another tweak to the differential phase splitter is to replace the common cathode resistor with a constant current sink (CCS).



The Web Site of Eric Kingsbury AKA Poindexter
 
PRR, without arguing against the notion that it never hurts (OK, rarely hurts) to have a bigger transformer than necessary, I'm not sure I follow you here. Plugging the values into PSUD, assuming a 33 ohm primary (the resistance of a transformer of about the right size that I have sitting nearby), 2 ohms total ESR for the cap, and a 1000 ohm load, I get an rms transformer current of 0.336A, which multiplied by the 220V rating leaves me safely under 100VA. Output voltage is 246V, close enough to yours (the difference is that I added the additional 100 ohm-100uF RC filter in the schematic).

Now, there may be an argument that the current spikes could saturate the core and overheat it, while spraying out a lot of hum. They do peak at a little over an amp, if one is to believe the sim. To see if that's true, I breadboarded the circuit (transformer on hand has slightly lower voltage, 200VCT, but my line voltage runs a little high) and saw something pretty close to what PSUD predicted, 232V. The transformer did not get particularly warm after a half hour of running, but the load power resistor was fairly toasty.

I wouldn't argue with stepping up the transformer VA rating a bit, but it would seem that 100W should suffice.

edit: looking at Morgan Jones's Bevois Valley circuit, I see that he's using a 300-0-300V@200mA power transformer to feed two channels from a cap input supply. If my math is right, that's about 120VA, not too far off target.
 
Thanks for all replies guys!
I discovered today that i can specialorder my HT transformers from a local store (EI-core), they can give me transformers up to 85VA rating, so i will build a dual-mono amp, with 85VA for each channel, that ends up in total 170VA for stereo, that should be sufficient for B+ dont you think?
I think i will order them with a 250VAC secondary.

The filament supply comes from a 2*9VAC transformer and will be regulated DC at the tubes.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.