Hoffman Amplifiers Tube Amplifier Forum
Amp Stuff => Tube Amp Building - Tweaks - Repairs => Topic started by: scstill on September 07, 2023, 11:16:12 am
-
I built an amp a year or so ago that I really liked, called Weekender as it was housed in a ladies makeup case. https://stillampd.com/weekender. Design is dual triode 6bl7 in push pull with dual triode 6sn7 as PI and another 6sn7 as two stage preamp. PT and OT from an old webster reel. Weber 8"
Decided to replicate the design but housed in old Philco radio cabinet so the layout was different including the PT, OT and Celestion speaker. Called the PhuelCel8 (Philco-Celesion-8"). Problem is that the new build is not as loud and breaks up much sooner than the original.
Double checked all voltages and bias and they were similar between builds (marked on Schematics). Double checked all resistors and caps and they are the same between builds. All wiring double checked. Swapped the tubes and the speakers between the builds. But the new amp still sounds as it did before not as loud and breaks up early.
The rectifiers are different (6x5 in original, 5y3 in new) but voltages are similar. OT are different but the power handling and impedances are similar (~14w, 8300:8ohm Original; 10w 8200:8ohm New).
What else might be going on here?
-
additional new build pics
-
Site cert expired 23 days ago. Chrome + malware detection will not allow comm with your site - You should fix that.
Regards,
--Pete
-
Although no NFB, I would still try swapping the OT leads.
And, did you try swapping the OTs between the amps? Since its the major difference I would go there.
Cool looking amp.
-
Probably the speaker you’re using
-
Probably the speaker you’re using
1st post states he swapped the speakers.
-
Although no NFB, I would still try swapping the OT leads.
And, did you try swapping the OTs between the amps? Since its the major difference I would go there.
I might try that...
Do you think I could simplify the experiment by swapping just the three input leads, grounding the two chassis' together and use the speaker jack from the swapped OT?
Although if that is the issue then I'm stuck with the way it is (the orig amp OT is somewhat of one-off find) and the replication build becomes just another different sounding amp.
-
BTW the original OT was driven by two 6v6s at low (215v) plate voltage. Thinking that would place the power at about 10-15w
Both plates of the 6bl7 are spec'd at 12w (the Hammond 125d at 10w was a compromise for fitment)
Relative to OT characteristics would the design be louder and cleaner if the OT was beefier say 20w?
-
Do you think I could simplify the experiment by swapping just the three input leads, grounding the two chassis' together and use the speaker jack from the swapped OT?
If I understand - yes I think so. but if you get a buzz (not the good kind) or hum (not the good kind) you may have created a ground loop.
About OTs - As a general rule, yes I think a larger OT will produce more clean volume. But its also more complicated than that. Valco, for example, is known for using undersized OTs - the perceived volume is pretty darn loud. Swapping is the acid (not the good kind) test.
-
Probably the speaker you’re using
1st post states he swapped the speakers.
Yes, and the Weber magnet and coil* may have been more efficient and the cone may have had different breakup characteristics (admittedly, this is a presumption as no details are provided about the design of that speaker)
*Weber vintage style speakers have a really narrow pole gap which makes them more efficient (magnet type and size being equal)
(But it could also be a difference in the OT reflected load, or amount of iron)
-
When I swapped the vintage ceramic Weber into the new design it still sounded less loud and early breakup.
When I swapped the ceramic Celestion into the original design it sounded loud and clean.
I figured the issue is not the speaker
-
Problem is that the new build is not as loud and breaks up much sooner than the original.
When I swapped the ceramic Celestion into the original design it sounded loud and clean.
what'd it do for the new build?
-
Putting the big weber in the new build did not help. It was still lower volume and earlier breakup than the Orig.
I also tried a 2x12 Jensen cabinet with similar results. Thinking it is likely caused by the OT or maybe even the PT. My surprise is that I tried to replicate the transformers Specs as close as possible yet there is a big sound difference. IN fact the Original PT is around 440vct @100ma where the New PT is 550 @220ma (why I have a dropper resistor before B1).
Gonna go thru the wiring one last time, if nothing, Then I suppose it is what it is. Sad but True.
-
Could be different dc resistance across the OT?
I wonder where the breakup is mainly happening. I'm currently learning (the hard way) how important the PI is to distortion and perceived volume, as well as how it interacts with the power tubes.
What would happen if you adjusted values at the PI to experiment?
-
I put the amp parts back to starting point and ran through many settings master low/high, tone low/high, vol low/high; to be honest it seems to always be in breakup. I cannot seem to get it to sound clean even with guitar vol rolled off. Something is very strange.
BTW - the orig OT impedance was measured with voltage in/out, it is 8344:8
the new OT was spec'd by Hammond at 8200:8
On the PI operation, relative to cathode, the grids are at -66v on the Orig and -70v on the new build. Plates are 115v on Orig and 117 on the New. Thinking that is close.
-
BTW - I have a new "stupid things you've done on a build" entry
I swapped the 6bx7 for the 6bl7. BUT at the design bias of ~-13v the 6bx7 draws 2x75ma vs 2x40ma for 6bl7
The first 680 dropper resistor was 3w which is fine for the 6bl7 but the 6bx7 needs 3.75w
So it smoked - but just a little :-)
Will get some 5w replacements
-
The 6sn7 looks like a monster tube. It's rated at 5Wish. So, it should be happy at 300v @ 15ma.
I see your load line for your PI here, and I'm probably wrong, but hopefully not too far off.
Even where it's at, you could probably double your cathode resistor and gain some clean headroom, but it isn't far from cutoff.
You could halve (or more) your plate resistor, and dial in the cathode to suit. Then you'd have to change the tail to an appropriate value, Merlin suggests 25-30 % of plate voltage.
Or maybe there's a tube that could fit into that circuit.
-
… I tried to replicate the transformers Specs as close as possible yet there is a big sound difference. IN fact the Original PT is around 440vct @100ma where the New PT is 550 @220ma (why I have a dropper resistor before B1).
…
Matching operating conditions at idle is only valid when the amp is idling. Try matching the HT voltages supplied to the OT CT etc when the amp is putting out full cranked power.
-
The LTPI is a bit out of whack. How it works well with the original may be related to the tube you're using. TubeCAD shows the values chosen result in a heavy output asymmetry. Those values are best used with much higher mu tubes such as the 12AT7/12AX7.
Suggest:
Ra1 - 47KΩ (inverting)
Ra2 - 56KΩ (non-inverting)
Rk - 3.3KΩ
Rtail - 15KΩ
Ib will be about 2.4mA so you may want to tweak the R value (reduce R of 2.7KΩ 2W) of the B+ rail that feeds the LTPI and other preamps to keep around 270VDC. With B+ at 270V, should swing about 70Vpk.
Simply changing Rk would probably make a world of difference - Balance is still out; 100KΩ Ra1 wants about 125KΩ for Ra2 to balance well with your chosen tail R - Stuffing a 3.3KΩ for Rk instead of a 470Ω, to start with you'd likely realize a significant improvement overall.
--Pete
-
based on good suggestions, I did some experiments on the PI to see if anything changed
1- I changed inverted and non-verted PI outputs as they were feeding different 6bl7 grids than the Original - no change
2 - I changed inverted plate resistor to 81k as PI Calc (below) suggested almost identical outputs - no change
3 - I changed Rk to 3.3k - no change
4 - I tried a few 6sl7 (amp factor = 70 vs 20 for 6sn7) - no change (maybe a little louder but definitely not clean)
Thinking it might be somewhere other than PI (but open to more thoughts).
With the volume set to 1 it starts to distort should be more like halfway.
-
That's too bad it didn't help out your issues. You may want to consider all of Dummyload's suggestions, IE. reducing both plate values, and also changing the tail resistor. Also, you could reduce the grid leaks, I believe they are at 1M right now? Merlin has a really good article on designing the LTP, I'm sure you've read it, but his info sure helps me when I'm stumped.
If you have a scope, or a listening amp you can probe around and see where it starts to get hairy.
-
I went through a few more experiments after suggestions here and re-reading Merlin LTP
1) Ra1 Ra2 = 35k for 200v (from 115v with 100k; HT is ~270)
2) Rt = 10.75k for 64.5v
3) Rb = 470ohm keeps the bias at -3v which is in the center of the load line (previously tried 3.3k)
Result was no change to the distortion. The distortion is not unpleasant, just frustrating that I cannot get any clean out of this amp.
So weird that two previous builds (weekender and westy) with same circuit sound really good.
https://stillampd.com/westy-02
https://stillampd.com/weekender
-
I had a closer look at both schema's side by side. Aside from the tilted LTPI, I missed the first RC stage in the main B+ rail before the power tube plate tap. That high source resistance (667Ω) is probably current starving the amp during big signal peaks. Try replacing that with a 47Ω, perhaps even a 22Ω. PSU DesignerII sim shows about 320V at B1 with 47Ω R, so shouldn't be an issue for use with 6BL7 triode. You'll want to swap out 470Ω under output tube cathodes to re-bias, the rest of the preamp ckts. won't care about the slightly higher B+. The 667Ω may work out as a replacement for the 470Ω.
Sim shows about 325V at B1 and just under 300V at B2.
--Pete
-
... first RC stage in the main B+ rail before the power tube plate tap. That high source resistance (667Ω) is probably current starving the amp during big signal peaks. Try replacing that with a 47Ω, perhaps even a 22Ω. ...
It's almost like high-Resistance was a reason that old amps used a choke in this spot. :icon_biggrin:
-
Try replacing that with a 47Ω, perhaps even a 22Ω.
Sim shows about 325V at B1 and just under 300V at B2
The reason that I added this 680 dropper resister was to match the B+ voltages in the "original" design.
The "new" design PT supported tubes (in its previous life) that would require 220ma which is much higher than "original" design PT (105ma) and that is needed in this amp design.
But I did try the suggestion dropping the resistor to 47ohm (resulted B1 of 310v and B2 of 290v).
It did not seem to make a difference as there is still no real clean. This was done as a quick experiment adding a parallel resistor to see if it made a difference so I left the 6bl7 bias resistor the same. I suppose I should make this a bit more permanent, but I wanted to see the positive improvement first.
What if I just used B1 for PI plates and took the 6bl7 voltage from the front side of the 680? would be 270v for Pre; 290 for PI and 315v for power tube. That is the most I will get from this PT. BUT, recall that the measured voltages of the "new" design as it currently is, closely match the "orig" without any changes yet there is no clean in the "new", even with amp volume at 1 or 2. Something odd.
-
If you lower or remove the dropping resistor at b+, then you should be able to dial in the cathode resistor at the output tubes for higher cathode voltage.
Looks like you're running about 15mA and 4 watts/tube.
Hard to tell from the datasheet, but it seems like their maximum watts should be around 10/side. It's weird, because it almost seems like on a push/pull you can increase it to 12W/side.
If you can increase your grid voltage at the power tubes, you should be able to increase your clean headroom, and still keep your plate voltages somewhat within spec.
-
... OT are different but the power handling and impedances are similar (~14w, 8300:8ohm Original; 10w 8200:8ohm New).
Can you take/post another photo that clearly shows the pins you're using on that Hammond 125D?
-
Hello,
I think V1 and V2 should be 6SL7 and not 6SN7.
With the 6SN7 they would be more than very hot biased.
-
I think V1 and V2 should be 6SL7 and not 6SN7.
With the 6SN7 they would be more than very hot biased.
Thanks. I am always playing around with those two tube types.
I tried it again with 6sl7 in both or either place and it is very hot with a lot more breakup
6sl7 gain is 70 (close to 12at7) 6sn7 gain is 20 (like 12au7).
BTW - when I roll the guitar volume to about half (lower than it should be) it is fairly clean up, then move the master up to get it loud. My clean test is typically a full strum chord (harder to get clean) verses individual notes
-
... OT are different but the power handling and impedances are similar (~14w, 8300:8ohm Original; 10w 8200:8ohm New).
Can you take/post another photo that clearly shows the pins you're using on that Hammond 125D?
Shown below. I have an on-on-on switch to chose 3 impedances. I use 2-5 for 8200:8ohm (which is the speaker I'm using) and 3-5 for 8800:4ohm; thinking I will need more switch poles to get 2-6 for 8000:15ohm. My switch schematic is also below.
Earlier I made a mistake and connected 5 and 6 together thinking that would allow me to get the 15ohm out but that shorted the coil and gave me a lower output (can see this in initial post). When I removed the short I got back to a decent volume (still too much breakup)
-
roll the guitar volume to about half
so your input signal is overdriving the preamp section. without eyes it's harder, but you can "bypass" stages. I always start with the TS, isolate it from the signal path, test.
IF NO;
put back, isolate the 1st gain stage from signal path
IF NO;
get a scope :icon_biggrin:
-
roll the guitar volume to about half
so your input signal is overdriving the preamp section.
Got me thinking....
The first preamp grid stopper of 33k was a lifted design (worked before) and is used by Fender into a 12ax7
When I calc for the 6sn7 using Merlin's formula I get 96K (http://www.valvewizard.co.uk/gridstopper.html), should also cut input signal, but he says the bigger the resistor the more hiss you get. Although some older Gibson amps used 100k into 6sj7 pentode and 6eu7 triode preamp. What do you think?
-
Changing the grid stopper will not cut the input signal level.
-
Earlier I made a mistake and connected 5 and 6 together thinking that would allow me to get the 15ohm out but that shorted the coil and gave me a lower output (can see this in initial post). When I removed the short I got back to a decent volume (still too much breakup)
This kind of reads like you may have fried some of the secondaries of the output transformer.
The marshall 18W would probably be a better choice if you want an OT with switchable secondaries.
--Pete
-
Earlier I made a mistake and connected 5 and 6 together thinking that would allow me to get the 15ohm out but that shorted the coil and gave me a lower output
This kind of reads like you may have fried some of the secondaries of the output transformer.
The marshall 18W would probably be a better choice if you want an OT with switchable secondaries.
I removed the 125D from the circuit and ran a voltage test (10v Primary)
For all of the valid pins at 8ohm (from Hammond chart) I got the correct impedance in each case.
I then ran two tests using spare OTs. One test with a 125E (15w) and another with a Princeton Reverb OT (15w). These sounded essentially the same as the 125D in the circuit. I might conclude that the OT was not damaged and the 125D OT is not the issue with the early (always) breakup.
These OTs are close enough to the suggested Marshall but unfortunately no way that OT would physically fit in this chassis or cabinet. I'm gonna get one though for future use (only $42 from Hawk)
-
Changing the grid stopper will not cut the input signal level.
Yes of coarse, grid is dead end, current doesn't flow.
Thanks for the reminder.
-
will not cut the input signal level.
this might;
1) unsolder the leads to BOTH grids
2) use a gator-clip to "jump" the now free leg of the 33k and the now empty V1b's G1
3) use 2nd gator-clip to ground V1a's G1
......
10 minutes later;
you're runn'n rifs, evaluating the change in sensitivity n pre-amp gain.
-
... I have an on-on-on switch to chose 3 impedances. ...
Earlier I made a mistake and connected 5 and 6 together ... When I removed the short I got back to a decent volume (still too much breakup)
You could now replace that complicated switch with a single-pole double-throw, for 2 un-ambigous wirings of 4Ω and 8Ω. It might even be worth bypassing the switch altogether, and staying with a single 8Ω output for testing (just to eliminate extra complicating factors during testing).
BTW - when I roll the guitar volume to about half (lower than it should be) it is fairly clean up, then move the master up to get it loud. My clean test is typically a full strum chord (harder to get clean) verses individual notes
Here's how I would have approached this issue from the beginning:
- Measure the DC volts at each stage (you've done this)
- Look at the bias voltage of each stage; often this is simply the DC volts at the cathode.
- Play the amp, slowly increasing the volume until distortion sets in (if that means rolling back the guitar volume, so be it).
- Measure the Peak AC Volts at each grid in the amp; annotate on your schematic where you also have the DC Volts.
- Compare the Peak AC Volts to the bias/DC volts for each stage. Any tube will surely distort when the peak of the applied AC volts equals or exceeds the tube's bias.
I bet this whole episode is as simple as over-driving some stage in the amp. Maybe you have a tube that is amplifying stronger than your original amp. Maybe something else is happening. Either way, measuring signal-level vs bias will likely point out where to focus your attention.
-
- Play the amp, slowly increasing the volume until distortion sets in (if that means rolling back the guitar volume, so be it).
- Measure the Peak AC Volts at each grid in the amp; annotate on your schematic where you also have the DC Volts.
- Compare the Peak AC Volts to the bias/DC volts for each stage. Any tube will surely distort when the peak of the applied AC volts equals or exceeds the tube's bias.
I bet this whole episode is as simple as over-driving some stage in the amp. Maybe you have a tube that is amplifying stronger than your original amp. Maybe something else is happening. Either way, measuring signal-level vs bias will likely point out where to focus your attention.
I sort of did this already but didn't write down the values (I repeated and wrote values below)
The first ac value is with guitar set to 7 (just starting to OD)
The 2nd ac value is with guitar set to 9 (in OD)
The Amp Volume was set rather low about 2 or 3
Not sure if there is a better way but I clipped AC meter to the grid and played guitar
I looked for the highest signal (usually on hard strums and quite fleeting) and wrote that down.
I don't see any issues (do you all?) The power stage shows OD at Guitar 9.
Recall that I replaced tubes from the "original" into the "new", so I decided early it was not a tube issue.
Maybe its just how this build is as it seems that we checked everything.
Just wish the "original" could be replicated or I could point to why its different
-
I sort of did this already but didn't write down the values (I repeated and wrote values below)
The first ac value is with guitar set to 7 (just starting to OD)
The 2nd ac value is with guitar set to 9 (in OD)
The Amp Volume was set rather low about 2 or 3
Thanks for this!
Power Tubes:
Bias: 14v
Light Dist: 10v (probably occasional peaks over that)
Strong Dist: 27v (tube is being slapped silly)
Note: The DC volts at the cathodes during the "Strong Distortion" signal levels almost certainly has shifted from 14v to something nearer the 27v peak grid signal. The bias-shift is normal for overdriven output tubes, and the near doubling of the signal level (27v peak grid-drive vs 14v idle bias) probably means your signal is half-wave rectified at the 6BL7 grids. Heavy distortion (literally "10 pounds of s@%t in a 5lb bag").
Phase Inverter:
Bias: 2.2v (68v - 65.8v on either side of the 470Ω resistor)
Light Dist: 0.6v (probably stays clean for all signals)
Strong Dist: 2v (almost certain distortion on peaks)
Note: Your measurements at the grids of the phase inverter are invalid, because they are "bootstrapped" (see "Input Impedance" here (http://www.valvewizard.co.uk/acltp.html)). That actual bias has to be measured as the difference between voltages on either side of the cathode-bias resistor (470Ω).
Gain:
V2B Smaller Signal: 10v (output tube grid) / 0.46v (V2B grid) = ~21.7x
V2B Larger Signal: 27v / 1.4v = 19.3x
V2A Smaller Signal: 8.5v (output tube grid) / 0.6v (V2B grid) = ~14.2x
V2A Larger Signal: 27v / 2v = 13.5x
Note: Gain to each output is unequal (expected with this type of inverter when using same-plate-load resistances) so the inverter always adds at least a trace of 2nd harmonic distortion. But notice how the gain seems to drop with higher signal level? A small amount of that is likely the inverter reaching its limits and squashing peaks; most of that is due to heavy grid-current at the output tubes clamping the output of the phase inverter. Stage-gain isn't dropping so much as it is being peak-limited because the power section is grossly overdriven.
Stage 2:
Stage 1:
Both have peak grid signals well below their bias voltage, so we probably don't need to worry about these stages distorting prematurely. That said, if the goal was to keep them clean then they are biased/loaded improperly. Each is biased so it has 2+ volts of input headroom in one direction, and more like 8 volts of input headroom in the other direction. See the location of the Blue Dot on the 6SN7 loadline at bottom.
If the goal is to "maximize the clean" for these 2 stages, then cathode resistors should be larger, plate current and/or plate load resistors smaller, and plate voltage higher. A good starting point is a plate voltage of 0.6x to 0.7x of supply voltage (which is a common rule of thumb for triodes).
That said, signal levels are typically low, so perhaps lack of center-biasing for these stages is not impacting things too much. It's hard to say without knowing the peak AC Volts output from the plate & before the Volume and Master Volume controls.
Bottom Line: This iteration of the circuit is flat-out pummeling the output tubes; it couldn't do anything but distort. Perhaps repeat the procedure with your original iteration to find out how they differ in signal-levels vs bias at each stage.
(https://i.imgur.com/SXMX15t.png)
-
Thanks HBP (and everyone) for great insight. Got me thinking that I should relook at the original design.
I went through a design calc for each stage (document below), using Merlin guide as I understand it.
Goal was to create more clean by increasing plate voltage and center biasing
Would appreciate any comments and corrections before I start cutting (measure twice cut once).
The plan is to go with the mods that are highlighted in each stage design.
-
You have 12AU7-type running with 12AX7 bias resistors. The 5:1 change of Mu suggests a similar change of bias resistors, or maybe somewhere between.
Eyeballing without ruler or calculator, I figure the 1.5K resistors could be 3k or 5k to suit the low-Mu tubes.
The plate resistors do NOT normally equal the nominal plate resistance; that throws-away half the gain and signal swing. Plate resistors are picked higher than tube internal resistance but lower than external load resistance. 100K is not wrong here.
-
The plate resistors do NOT normally equal the nominal plate resistance; that throws-away half the gain and signal swing. Plate resistors are picked higher than tube internal resistance but lower than external load resistance. 100K is not wrong here.
Isn't the internal plate resistance 7.7k for 6SN7 and 12AU7?
Everyone in this thread has been saying decrease the plate resistors to get more plate voltage and more cleans. This is the first time in the thread that 100k was considered acceptable.
We all agree that higher cathode resistor give more clean.
At 9ma typical for 6SN7 at 270v supply doesn't that say 30k for plate resistor? 26k for 12AU7 10.5ma.
Am I missing something?
Two load lines below for the 6SN7 one with 100k and one with 30k. Several bias voltages shown.
Are my calculations ok? Which of the color dots will provide a cleaner output? Its also been suggested that we should be center biasing so maybe the green on the red curve?
BTW - the green dot on the blue curve is the current design Rplate 100k and Rbias 1.5k, but I do not get 6v bias, I measure something more like 3v
-
I do not know why your two "identical" amps are different.
As for how to use 6SN7 in a guitar amp (not a popular choice), you can experiment (resistors are cheap), take a vote on social media, or look at the publications of tube factories and their workers who hoped to earn a living from happy tube designs.
We see 100k is a medium load for 6SN7 in a resistance-coupled amp. Operration at 268V will be very-near, about 90%, the 300V supply data. Since your actual loads are nominal 1Meg, use the "0.47"(Meg) condition. Cathode resistor can be very near 4k.
-
Is it not beneficial with an LTP phase inverter to increase voltage at the plate, therefore allowing higher voltage at the cathode? Does the elevated cathode not give you more voltage swing? Not trying to be argumentative, just curious.
-
voltage swing
the "swing" is ~~~~ difference between plate VDC and Cathode VDC
so "widen" that range, get more swing.
The chart PRR referenced, the 2nd to last column, is Eo, a ballpark # for "swing"
-
voltage swing
the "swing" is ~~~~ difference between plate VDC and Cathode VDC
so "widen" that range, get more swing.
The chart PRR referenced, the 2nd to last column, is Eo, a ballpark # for "swing"
Ok, so I guess my next question would be... Why elevate the cathode at all? I assumed it was to allow for negative signal voltage as well as positive.
-
you want both positive n negative halves of the signal to have "equal" range to swing. the closer you have one side to zero or B+ the more asymmetrical the swing become
-
And I thought I was crazy chasing tail.
-
Is it not beneficial with an LTP phase inverter to increase voltage at the plate, therefore allowing higher voltage at the cathode? Does the elevated cathode not give you more voltage swing? ...
No.
Ok, so I guess my next question would be... Why elevate the cathode at all? I assumed it was to allow for negative signal voltage as well as positive.
"Voltage across resistor" ("tail resistor" from cathode to ground (https://www.aikenamps.com/index.php/designing-long-tail-pairs-the-load-line-approach)) kinda-sorta help force equal-but-opposite current in each section of the long-tail pair. It doesn't do a great job, but the basic idea is to fake a constant-current source (though a real, electronic constant-current source is very much more effective).
- We drive signal into one side of the long-tail.
- We rely on the "fake constant-current source" and "cathode-drive" to drive signal into the other side of the long-tail.
- If it all works perfectly, we get equal-and-opposite outputs.
- It doesn't work perfectly, so we trim one side's plate load smaller than the other side to make up the difference.
-
Ok, so rather than do some tree work this past weekend, I was musing about this amp - What are we hearing and what does it really sound like with the lower mu cascaded gain stages. I have a breadboard that was unused, so I cobbled together something similarto your Weekender. I used SSR and have a PT that got me close enough to 300V B+ with SSR. I used a NOS Stancor A-3830 15W output transformer strapped for 8Ωvc:10KΩa-a.
Other changes were bias points for V1 and V2, I left V3 (6BL7) relatively same. Initial impression was it sounds decent with the 6SN7, but kind of anemic with single coil pickups (Fender Strat w/ Seymour Duncan Antiquity), and I omitted the Master Vol.. I used 3.3KΩ cathode resistors which were a better match for the 100KΩ plate load with 280V B+ with the 6SN7 - Note that the 3.3KΩ cathode resistors also work well with 6SL7. In fact with 6SL7 in V1 and V2 it woke up with much more positive results. The entire amp with 6SL7/6BL7 draws about 37mA.
I have stash of 6DZ7 - I plan to replace the 6BL7 with one of those just to satisfy my curiosity - More on that later. In the meantime I attached a schema of what I built - Link to pics as well.
--Pete
https://www.dropbox.com/scl/fi/vfnkp58428jj38ln1d3ph/20230925_015324.jpg?rlkey=ym5y4111j5btbjcxxjfjtc683&dl=0
NOTE: Schematic is wrong - States V2 is 6SN7GTB, it is actually a 6SL7GT and the telemetry indicated reflects that.
-
I also have been busy...
I redesigned each stage of the PhuelCel8 (Philco Celestion 8") so that they are balanced (ie equal voltage swing) which I understand will provide a proper clean headroom. I discovered that there were some issues with the balance of the original design, maybe why it was "ODing". This would include my earlier amps Weekender, Westy as well as the Moonlight (someone else's design http://www.frontiernet.net/~jff/moonlight_bl7/Moonlight_AllOctalDoublePower.html).
But the Weekender and Westy sound good so they are staying for now, https://stillampd.com/westy-02
Remember my complaint was that the amp had no clean headroom. Seemingly always in overdrive.
I am pleased to say that the new design is great. It behaves as I think it should. Clean until about half volume then a nice crunchy overdrive increasing as the vol is rolled higher. To me it sounds good. I still can't figure out why Weekender and Westy sound good with the original design (lifted from Moonlight). I will create and post the PhuelCel8 design picture essay to stillamp'd site soon.
Changes include;
1) Change the plate resistor for each 6SN7 from 100k to 30K
2) Change Cathode resistor on LTP from 470 to 800
3) Change Tail resistor on LTP from 22k to 10k
4) Change the Cathode Bias resistor on 6BL7 from 470 to 360
5) Get supply voltage for 6BL7 from first node (rather than 2nd)
6) Get supply voltage for 6SN7 LTP from second node (rather than 3rd)
The revised schematic is provided below
in addition for those interested, I have included the load line calculations for each stage showing design before (unbalanced) and after (balanced). Appreciate any comments as this is the first time through this design for me. Much easier just to copy someone else's design I suppose.
-
I used a NOS Stancor A-3830 15W output transformer strapped for 8Ωvc:10KΩa-a.
Other changes were bias points for V1 and V2, I left V3 (6BL7) relatively same. Initial impression was it sounds decent with the 6SN7, but kind of anemic with single coil pickups (Fender Strat w/ Seymour Duncan Antiquity), and I omitted the Master Vol.. I used 3.3KΩ cathode resistors which were a better match for the 100KΩ plate load with 280V B+ with the 6SN7 - Note that the 3.3KΩ cathode resistors also work well with 6SL7. In fact with 6SL7 in V1 and V2 it woke up with much more positive results. The entire amp with 6SL7/6BL7 draws about 37mA.
I have stash of 6DZ7 - I plan to replace the 6BL7 with one of those just to satisfy my curiosity - More on that later. In the meantime I attached a schema of what I built - Link to pics as well.
10k load for the 6BL7 is also on my radar to try (my OT has multiple taps). It will lower the load line slope (from my 8.8k design) and should stay balanced with same cathode resistor. BTW - I had another "stupid things you do" entry as initially I thought I had an 8ohm speaker and designed for 8800:8 haha; it was 4ohm so I really had 4400 which also contributed to some of my issues (load line was extremely sloped, 263ma class B for 290v!!). Of coarse it has been corrected.
3.6k cathode resistor for 100k is what I calc (orange dot on pre amp load line) so 3.3k is a good choice. But the load line was so low and so flat. That is why I went with 30k plate resistors.
Thanks for the 6DZ7 intro, did not know about it, but I want to also try this dual pentode in a future project.
-
Please excuse the threadjack - The following evolved from the experiments of the mockup of your amp. I have a small stash of the 6DZ7 and have always wanted to fool around with one or a pair in a P-P arrangement. I was pleasantly surprised with the results - Should have fiddled these long ago.
Spec sheet says about 12W with roughly same conditions presented.
https://frank.pocnet.net/sheets/093/6/6DZ7.pdf
Schama attached. Link below to the experiment.
https://www.dropbox.com/sh/3tkvhtgpsb6wcdo/AABl8SOl0wDfI9EShnGAXCWba?dl=0
Regards,
--Pete