A bit pedantic, perhaps. These are direct coupled amplifiers, but not really "Loftin/White".
The original here: Loftin White: September 2007
It was based on a pentode input tube.
There's a simple schematic of a triode version, halfway down the page.
Note the connection from the top of the output cathode resistor, to the top of the Unbypassed input cathode resistor. That's the crucial difference, and it cancels power supply noise. I've built a 6SN7/801 version of this. It's quiet as a grave and uses only film caps throughout because it doesn't require a lot of PS filtering. Took me a while to get my head wrapped around the concept.
Back in their day, caps were large and expensive - relative to their capacitance. Note that the original version uses only a half-wave rectifier.
Sheldon
The original here: Loftin White: September 2007
It was based on a pentode input tube.
There's a simple schematic of a triode version, halfway down the page.
Note the connection from the top of the output cathode resistor, to the top of the Unbypassed input cathode resistor. That's the crucial difference, and it cancels power supply noise. I've built a 6SN7/801 version of this. It's quiet as a grave and uses only film caps throughout because it doesn't require a lot of PS filtering. Took me a while to get my head wrapped around the concept.
Back in their day, caps were large and expensive - relative to their capacitance. Note that the original version uses only a half-wave rectifier.
Sheldon
Last edited:
The one on the right embodies the LW principles. The green line represents the output AC current loop.
Attachments
Last edited:
So I noticed a hum pretty quickly and just had time to troubleshoot it today. At first I thought it was heater realated (I’d forgotten to provide a ground reference) but it turns out to be power supply noise. There’s about 10mV of ripple at the speaker terminals. To mitigate it I’ll add a noise cancelling cap from b+ to cathode but I’ll need to wait for a parts shipment to do it. Beyond that I tweaked the values a bit so I have 90% dissipation with 375v across the power tubes. Output power seems to be about 7W peak. I also tried bypassing the e88cc cathodes but I didn’t hear a difference so then I added 6db of cathode feedback. The sound cleaned up quite a bit but I think I like the more “live” character of leaving it out. I’ll probably build a switch for it.
Next update: I swapped the 100uF cathode bypass caps for 40uF film and added 4uF caps on each channel from B+ to the cathode for noise cancelling. The cancellation is quite effective though I can still hear the hum if I get very close to the driver, overall I consider it successfull. I used a film for the large value because the tolerances are tighter. I probably should have gone with a 50 and a 4.7uF the better match the mu of 10.5 but started getting spooked over the size/price.
I think the prototype is done and I now need to turn my attention to a proper enclosure.
I think the prototype is done and I now need to turn my attention to a proper enclosure.
Attachments
Just a guess on my part, but I think you'll find the hum will decrease once you re-build the amp on the final chassis and shorten up the wiring.
Interesting circuit; thanks for posting!
I actually prototyped the TubeCad LW Auto Bias idea, low power version though.
Tubes are just *one* compactron 6FM7 dissimilar triodes per channel. That's a veritable SPUD.
And - yak - a pnp transistor instead of expensive 300H+ choke.
DC bias stabilisation comes in for free.
Zout is 0.5 ohm but only 2W max.
Z out is 1 ohm with half the impedance ratio in the OPT at 3W but THD doubles.
Attachments
As an additional thought I have wondered if the transition into clipping would sound different between the ccs loaded driver and the inductor loaded version. I’d imaging the ccs would hard clip a little below the power tube’s cathode voltage; the inductor would try to swing above the cathode but be clamped by the power tube grids.
Sorento - do you know how many volts your ccs drops when properly set up? I bought the parts for a ccs version but didn’t attempt it yet.
Sorento - do you know how many volts your ccs drops when properly set up? I bought the parts for a ccs version but didn’t attempt it yet.
There's 60 volts across the transistor at idle. Properly setup ... hmm ... don't know.
And yes, clipping of the CCS is hard.
In my case the feeble "power" triode is cut off long before the CCS, so the other half of a sine wave gets clipped first anyway ...
I was taking some power/loudness measurements today to figure out what my next amp project will be and thought I'd post some values in case a future DIY'er want to follow along:
EL34 [V]
Plate 520
Screen 518
Grid 90.4
Cathode 121.4 (50mA)
E88CC [V]
Plate 91.9
Cathode 2.05 (8mA)
Measured at the speaker on the 4 ohm tap, the amp clips visibly at 7V peak or 6.1W. It starts to sound congested above ~4W.
EL34 [V]
Plate 520
Screen 518
Grid 90.4
Cathode 121.4 (50mA)
E88CC [V]
Plate 91.9
Cathode 2.05 (8mA)
Measured at the speaker on the 4 ohm tap, the amp clips visibly at 7V peak or 6.1W. It starts to sound congested above ~4W.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.