18W 1958 / 1973 / 1974 Original Schematic

[geoff 1965]

found it colossal!
you can clearly see the 91R and specs of the dagnall
viewtopic.php?f=4&t=25508&p=244206&hili ... ue#p244206

[colossal]

found it colossal!
you can clearly see the 91R and specs of the dagnall
viewtopic.php?f=4&t=25508&p=244206&hili ... ue#p244206

Ah yes, I remember that one now. Thanks for finding that.

I found another photo of an original '67 and it clearly shows 100R as the bias resistor. Also of note is the original "Hygrade" Radiospares power transformer. There are two references I can find in the catalog;

300-0-300V, 120mA
6.3V, 3A centertapped
0-5V, 2A, 0-6.3V, 1A

and

275-0-275V, 100mA
6.3V, 2.5A (not centertapped)
0-5V, 2A - 6.3V, 1A

I am led to believe, however, that the 300-0-300 120mA Radiospares PT was used in the originals, even though the Dagnall 275-0-275 is used in the reissues.

Hygrade RS Mains.png

RS Hygrade.png

1974 circuit.jpg

[geoff 1965]

look at the B+ at the first cap on those old schematics colossal,325VDC,we know even with the old lower wall voltages that a 300-0-300 HT is nowhere near that!
i think the very first ones will have used the WEM PT's, however did RS make a 280-0-280 in that catalogue? i'm sure i saw one recently on ebay.

[colossal]

look at the B+ at the first cap on those old schematics colossal,325VDC,we know even with the old lower wall voltages that a 300-0-300 HT is nowhere near that!
i think the very first ones will have used the WEM PT's, however did RS make a 280-0-280 in that catalogue? i'm sure i saw one recently on ebay.

There is a 275-0-275, 100mA listed in the 1967 catalog. However, it does not show the heater filament supply as being centertapped, where the 300-0-300 filament supply is labeled as centertapped. In the photos above, specifically the one of that original amp, the solder tabs on the transformer match the 300VAC Hygrade PT. Mercury offers two Radiospares clone PTs for 18W amps; one at 301-0-301 120mA and the other at 316. Those are unloaded voltages. I use the former and mine sags to about 289VAC under load. The Mercury is wound for 120VAC input, some compensation for today's higher voltages where the original RS was 0-100-115-125 (US spec) and 0-205-225-245 (UK spec). An EZ81's rectification factor under load is roughly 1.15 x VAC. The ra for an EZ81 is 200R, a hefty impedance, at 300VAC input. An 18W pulling 87mA, 2x40mA for 2xEL84 plus a handful for the preamp and some transformer losses gives a 325-330VDC B+:

Mullard EZ81.png

I built a power supply design model in Excel based on the work by:

Schade, O. H., Proceedings of the I.R.E, 31, p341-361, (1943)
and
Reich, Herbert J, "Theory and Applications of Electron Tubes", 2nd Ed., McGraw-Hill, 1944

...and plugging in the numbers for the EZ81 and some guesses about the series resistance of the RS transformer, it estimated 326VDC as the rectified voltage. I did not massage the numbers to fit my argument, but just shotgunned in a few assumptions to see if it was even reasonable. The procedure is rather involved and I did not fit the curves for an EZ81. I would need to do this, plus have some real impedance numbers from the actual transformer to make a real prediction.

Colin's reference schematic shows 325VDC B+ and the PT at 300-0-300, with a 125R bias resistor.

storage/18w%20Schematic.gif

I am again wondering if Marshall just used what was available, which might explain why some amps had 91R, 100R; these might have gotten a 275-0-275 PT, and others got 125R and perhaps the 300-0-300 PT. Just speculating here, of course.

[crgfrench]

The last 100W 1959 I built was to the Friedman verified '68 Van Halen spec with 50-choke-32-100-16-16 and that amp was great.

Could you share the schematic/layout for that amp?
I'm thinking of doing something like that next.

[geoff 1965]

I’ve just swapped to the reissue values of 470uf on the el84’s and 22uf on the normal input cathode and it sounds awesome!
Keeping away from the 91/100R bias though,160R is giving me 11.5 watts PD and 12.4V on the cathodes with the 275-0-275 HT.

F7FFCE3A-35D1-42BB-921A-C4A0E27B3A74.jpeg

[colossal]

I’ve just swapped to the reissue values of 470uf on the el84’s and 22uf on the normal input cathode and it sounds awesome!
Keeping away from the 91/100R bias though,160R is giving me 11.5 watts PD and 12.4V on the cathodes with the 275-0-275 HT.

Was your Normal channel 820R/50uF before? What was the bypass cap on your power amp before?

My amp is 150R and 11.8W of dissipation per tube with 301-0-301 (unloaded) and a Mullard EZ81.

[geoff 1965]

I had 100uf on the el84’s and originally 47uf/820 on V1 when it was lite2 but i’ve separated the triodes now,22uf/1K5 and .68uf/2K7.

[colossal]

I had 100uf on the el84’s and originally 47uf/820 on V1 when it was lite2 but i’ve separated the triodes now,22uf/1K5 and .68uf/2K7.

Very good.

Are you still running V1 as a common anode mixer, so, Pins 1 and 6 are combined, but with separate bias 22uF/1k5 and .68uF/2k7 on pins 3 and 8? Or are you cascading V1a into V1b?

[geoff 1965]

Separate plate resistors colossal and yes cascading,I did a lot of experimenting with the parallel/cascade mod and got some good combinations but was’nt 100% happy with it.

[colossal]

Separate plate resistors colossal and yes cascading,I did a lot of experimenting with the parallel/cascade mod and got some good combinations but was’nt 100% happy with it.

Ah, ok. I did similar experiments with a cascaded channel, but still keep coming back to the parallel triode Normal channel. Both sound very good, but I really like that single paralleled triode gain stage.