Summer Holiday Fun build idea - comments please

[tubeswell]

I have posted this over at ampage just now, but thought I might have some luck getting feedback from the good folks here as well.

I have a PT with 240-0-240 100mA and 6.3V 2A and a fairly chunky 25Wish SE gapped OT 8k Pr with 4, 8 and 16R secondaries, plus the rest of the bits I need to get this rolling.

I've fancied a parallel SE amp for some time and also like the idea of an EF86 with Merlin's triode/pentode morph control. I also dig trem and have a spare triode and a pile of IRF820s so...

Here is the 1st cut at the schematic. Any suggestions/comments/criticism feedback is welcomed please. (The voltages are just guesses)

I am particularly interested in whether anyone here has any reckoning about adding NFB (none at present on the schematic), and also the relative sizes of the trem depth pot and the output stage grid load resistor (highlighted in Blue) - I was wondering whether a 100k pot and a 180k grid resistor would be more appropriate, or maybe a 100k pot and a 220k resistor?

TIA

[jojokeo]

Looks like a fun project & build.
For NFB, try adding a small resistor under that 12ax7 following the EF86 like a Fender Champ/Bronco w/ 47R and using 2.7k or better yet, get a 5k pot and adjust it on whatever tap you want to use to find your sweet spot, measure it and put that value in it's place. (thank Geezer for that idea not too long ago). Maybe put it on a switch too?
I've played plenty w/ EF86s and 5879s and I'm not big on the morph triode/pentode control. It may sound seem cool but in reality all it does is change gain of the tube. If it was me, I'd install a pentode/triode switch for that and adjust your gain on either setting w/ a variable cathode bypass cap control. This is a must have when using an EF86 especially, IMHO and really gets the most versitility from it.
Your last resistor values on the trem & output stage may likely need to be experimented with once the amp is up and running to find your happy zone?

Joe

[tubeswell]

Finished the chassis yesterday. (High Res) Pics here: http://music-electronics-forum.com/t23443/

[jojokeo]

Nice looking work tubeswell. There's a biggie mistake on the EF86 if you wired it up per schematic? pin 8 only goes to pin 3 - not to the grid of pin 9. It won't work that way, you're grounding out the grid effectively. Also w/ your values you may want to try subbing in a 1k5 cathode resistor instead of the 630r. Get back w/ your results on how it all sounds and works, especially the "morph" control function & if the trem is what you were looking for?

[sluckey]

That vertical line from pin 8 simply crosses over (or under) the horizontal line from pin 9 without connecting. I believe tubeswell would have used a dot at that intersection to indicate a connection.

[jojokeo]

Did you look at his last/latest link? That's exactly what he did.

[sluckey]

You're right. I had just looked at the schematic he posted here. Sorry, not trying to bust your chops.   

[DummyLoad]

sim tells a slightly different story on the PS concept. sure you want to put the screens after the 10K? possibly going to buzz a lot. your not using a vacuum recto, so you could be a a bit more generous w/ filter. it's SE, so it won't mind at all and IMO w/ SS and a moderate amount of filtering, you don't sacrifice tone. screen currents are an estimate and conservative and step @ 3seconds in both sims - from 4ma to 10ma full power. i also made an estimate on the plate load.

it is just simulation so your mileage may vary + i haven't had my morning coffee yet.    


-ISO

[jojokeo]

I'd like to have a better understanding of that program and interpreting the results. I need a teacher or a good help menu to get started w/ it.

[tubeswell]

Did you look at his last/latest link? That's exactly what he did.

  Bloody stupid drawing program  Fixed

[jojokeo]

I looked at your pic w/ the tube sockets but it was still hard to tell if you did so or not. Did you fire it up and get it playing?

[tubeswell]

sim tells a slightly different story on the PS concept. sure you want to put the screens after the 10K? possibly going to buzz a lot. your not using a vacuum recto, so you could be a a bit more generous w/ filter. it's SE, so it won't mind at all and IMO w/ SS and a moderate amount of filtering, you don't sacrifice tone. screen currents are an estimate and conservative and step @ 3seconds in both sims - from 4ma to 10ma full power. i also made an estimate on the plate load.

it is just simulation so your mileage may vary + i haven't had my morning coffee yet.    

-ISO

Thanks for the thoughts ISO.  I changed the PT and went with a vacuum rectifier. The schematic is as per above. I could try the screens in different positions. In present configuration, the CLC filter after the rectifier is a Mojo 777 (100mA 100R 3.5H) with a 20uF each side of it. Then there is a 20k/22uF then a 22k/10uF.  

On your graphs I take it the green band is the combined Plate & Screen current.  But what is the story with the screen current in the red and white bands? And what is the difference between those and the purple and grey lines?

Similarly with those screen current bands/lines and the plate blue and yellow plate current lines, why do all sets of lines dip sharply in the middle?

[tubeswell]

I looked at your pic w/ the tube sockets but it was still hard to tell if you did so or not. Did you fire it up and get it playing?

Not quite yet.  The EF86 is wired correctly with Pin 3 and pin 8 connected (see the corrected schematic I posted just before). That dot on the earlier schematic was an auto-idiosuncracy of the drawing program.

I have one more query about the PT. I noted that the case of the Weber PT I installed has no continuity with the chassis. The PT bolts have continuity, but they are isolated from the PT case with Nylon grommets. Steve Connor from over at MEF has suggested this might be unsavoury (but he could just be pulling my chain). Another amp tech I know locally (Clarry Schollum in Auckland) has said that is not unusual for some US-made PTs cases not to have any continuity with the chassis, and he had a Mesa Boogie on the bench yesterday when I was talking to him, and he tested it and it's PT case also didn't have continuity with the chassis. Any other opinions?

[DummyLoad]

the first graph - after 3 seconds you go to full power.... screen current increases - I increases; V decreases - plate current not so much. your B+ to screens and trem. ckt. transitions from 240V to 150V from idle to full power. B+ to EF86 and driver transitions from 220V to 130V. keep in mind the values i used for screen current i believe to be conservative.

green = B+ to plates.
yellow = B+ to screens and tremelo
blue = B+ to EF86 & driver

red = I thru 10K resistor
white = V across 10K resistor

grey = I thru 22K resistor
violet = V across 22 resistor

notice that V and I of R 22K fall to 0 for about 100-150mS - that can't be good.

second graph:

blue = B+ to plates
white = B+ to screens
grey = B+ to trem ckt.

violet = V across 500Ω
red = I thru 500Ω

teal = I thru 10K
green = V across 10K

yellow = I thru 22K
blue = V across 22K

forgot to check V for C(4) of 22K. sorri. assume B+ is ~280V  or V across 22K minus B+ to trem ckt. (grey line).

[jojokeo]

I have one more query about the PT. I noted that the case of the Weber PT I installed has no continuity with the chassis. The PT bolts have continuity, but they are isolated from the PT case with Nylon grommets. Steve Connor from over at MEF has suggested this might be unsavoury (but he could just be pulling my chain). Another amp tech I know locally (Clarry Schollum in Auckland) has said that is not unusual for some US-made PTs cases not to have any continuity with the chassis, and he had a Mesa Boogie on the bench yesterday when I was talking to him, and he tested it and it's PT case also didn't have continuity with the chassis. Any other opinions?

I recall a thread maybe 4 weeks ago where someone was having noise issues w/ his amp. He mentioned the tranny was older and possibly from a different or older build. Anyway, his PT was thought to be missing the fiber/nylon washers for isolation and that the PT coupled noise into the chassis w/out these. I've taken notice of this on a number of PT's I've seen since and it appears that most of them have them. BUT, they seem to be located only on one end or side of the tranny! This may be enough to isolate it properly? Here's some examples of Hammond's that were easy to show you what I'm describing.

[tubeswell]

I have one more query about the PT. I noted that the case of the Weber PT I installed has no continuity with the chassis. The PT bolts have continuity, but they are isolated from the PT case with Nylon grommets. Steve Connor from over at MEF has suggested this might be unsavoury (but he could just be pulling my chain). Another amp tech I know locally (Clarry Schollum in Auckland) has said that is not unusual for some US-made PTs cases not to have any continuity with the chassis, and he had a Mesa Boogie on the bench yesterday when I was talking to him, and he tested it and it's PT case also didn't have continuity with the chassis. Any other opinions?

I recall a thread maybe 4 weeks ago where someone was having noise issues w/ his amp. He mentioned the tranny was older and possibly from a different or older build. Anyway, his PT was thought to be missing the fiber/nylon washers for isolation and that the PT coupled noise into the chassis w/out these. I've taken notice of this on a number of PT's I've seen since and it appears that most of them have them. BUT, they seem to be located only on one end or side of the tranny! This may be enough to isolate it properly? Here's some examples of Hammond's that were easy to show you what I'm describing.

Cool Thanks

On first power up the amp is very quiet only a smallest hiss (touch wood) (Not even any NFB screeching)

[tubeswell]

the first graph - after 3 seconds you go to full power.... screen current increases - I increases; V decreases - plate current not so much. your B+ to screens and trem. ckt. transitions from 240V to 150V from idle to full power. B+ to EF86 and driver transitions from 220V to 130V. keep in mind the values i used for screen current i believe to be conservative.

green = B+ to plates.
yellow = B+ to screens and tremelo
blue = B+ to EF86 & driver

red = I thru 10K resistor
white = V across 10K resistor

grey = I thru 22K resistor
violet = V across 22 resistor

Thank you for being so helpful

notice that V and I of R 22K fall to 0 for about 100-150mS - that can't be good.

Why is that do you think? How can there be no voltage/current through the supply resistor?

second graph:

blue = B+ to plates
white = B+ to screens
grey = B+ to trem ckt.

violet = V across 500Ω
red = I thru 500Ω

teal = I thru 10K
green = V across 10K

yellow = I thru 22K
blue = V across 22K

forgot to check V for C(4) of 22K. sorri. assume B+ is ~280V  or V across 22K minus B+ to trem ckt. (grey line).

Cheers. Ill post the voltages in a minute - just taking them now

[DummyLoad]

I changed the PT

to ?

--ISO

[tubeswell]

I changed the PT

to ?

--ISO

Weber W025130EU (the 340-0-340 HT is hooked up, and I am trying different tube rectifiers)

[tubeswell]

PT No Load
HT 719VAC (359.5-0-359.5)
Heater – 6.79VAC
Rectifier – 5.27VAC

Loaded
5Y3GT, 6V6S (8K Pr Z), 12AX7, EF86

HT voltages
B+ = 415 VDC at reservoir cap (470 peak at startup)
10k node = 343 VDC (436 peak at startup)
22k node = 280 VDC (369 peak at startup)

6V6S
Plate = 406 VDC
Screen = 340 VDC 560R g2 resistor dropping 3V = 0.005A
Cathode = 19.8 VDC (across 472R measured) = .0419A

.0419A Ia+Ig2 x 386.2V a-k = 16W (minus 5.3mA Ig2 = 14W plate dissipation)

12AX7
Driver Stage (280 HT)
Plate = 182 VDC
Cathode = 1.31 VDC

LFO Stage (343 HT)
Plate = 208 VDC
Cathode =  1.94VDC

EF86
Plate = 120 VDC
Screen = 91 VDC
Cathode = 1.3 VDC

However when I plugged it in there was a problem. The guitar sounded really weak and muddy with hardly any volume, and then there was a spark above the 6V6 socket (when I was playing - it didn't spark at idle). When I switched it off (immediately) the 6V6S was really hot. Having a look now.

[tubeswell]

Well after that kerfuffle, (insanity break caused by troubleshooting mis-wiring and extending the height of the board off the chassis to avoid unwanted shorts), I figured out that the way I have the trem set up must be interferring with the 6V6 bias something wicked, so back to the drawing board for that part - looking for ideas please (maybe pre-amp driver tube trem wiggle??).

Anyhow I'm getting past myself.

I disconnected the trem and sorted out the NFB polarity and decided to swap over to the 'low power' HT taps on the PT. Still at this stage with one 6V6 plugged in running off a 5Y3GT. I also 'upped the 22k resistor to 27k, but I might put that 22k back again.

Anyway my first impression are that the 'Mojo' control is really effective at going from pentode to triode mode, and it has a wider range of tones and sounds than I wasn't expecting from a tweed princetony style amp - what with the 'Mojo' control and the volume and tone, there's quite a lot of scope with this little baby. And its loud - I couldn't turn it up past 4 with the little 15W alnico something speaker I found in the wardrobe. Scoops first prize. I will wire up the Greenback G10 later and try out the 2nd 6V6, although I have a hunch I'l be going to a 5U4G or GZ34  or going to those high taps.

I must get some soundbytes going.

PT
HT 558VAC (279-0-279)
Heater – 6.79VAC
Rectifier – 5.27VAC


Loaded
5Y3GT, 6V6S (8K Pr Z), 12AX7, EF86

HT voltages
B+ = 326 VDC at reservoir cap
10k node = 275 VDC
27k node = 209 VDC

6V6S
Plate = 321 VDC
Screen = 272 VDC 560R g2 resistor dropping 3V = 0.005A
Cathode = 16.4 VDC (across 472R measured) = .0347A

.0347A Ia+Ig2 x 304.6V a-k = 10.6W (minus 5.3mA Ig2 = 9.5W plate dissipation)

12AX7

Driver Stage (209 HT)
Plate = 127 VDC
Cathode = 1.23 VDC

LFO Stage (343 HT)
Plate = 155 VDC
Cathode =  1.75 VDC

EF86
(209 HT)
Plate = 92 VDC
Screen = 69 VDC
Cathode = 0.99 VDC

[tubeswell]

Okay, what I am thinking now is to aim to drive the output tube grid into cut-off by making the trem swing into really -ve voltage. (The PT has a 45VAC bias tap). I don't know if this will work, but please bare with me, and please please do chime in if you know how it could be done a better way etc.

Anyway the theory is that at present the grid sits at 0V (being a cathode-biased output tube'n'all), and I think that if I design the trem to swing only in the negative-voltage range, with a concomittent -ve bias adjustment on a dual-ganged pot (that is the same pot as the trem depth control), that I could sort of match the amount of -ve voltage required at the trem voltage swing mid-point to be set relative to the trem depth, so that the trem's most-positive swing was always at (or near enough to) 0V, and the output tube was driven into really cold bias, so that the trem doesn't adversely affect the normal biasing of the output stage (as it otherwise appears to be doing after I bench-tested a prototype of the previously dreamed-up incarnation). Please refer to the concept schematic* attached. TIA

* in the concept schematic the light-grey shaded line is meant to represent the dual-ganged pot, which in theory when turned CW both increases the trem depth and increases the -ve voltage of the trem swing mid-point)

[DummyLoad]

since your going fixed bias and all... why not just copy tried, true, tested and lovely sounding fender plan? princeton reverb AA1164 - no rules say you have use P-P output stage. pitch the sand state crap and use the 1/2AX7 in similar fashion. some tweaks on the divider for the -bias for your SE conditions and your good to go. be careful to not get too much wobble - it could make things unhappy quickly.

it looks like your plan for -ve adj. runs power tubes down closer to cut-off.

http://www.el34world.com/charts/Schematics/files/fender/PRINCETON_REVERB_AA1164.pdf

--ISO 

[tubeswell]

Hi ISO, I wasn't thinking about making if fixed bias really, rather I was just thinking about using the bias tap to power a variable negative voltage supply to set the trem swing's midpoint -ve w.r.t the output tube grids, in such a way that the most-positive part of the trem swing is at-or-around 0V (no-matter how much you crank the trem depth control).

[tubeswell]

Epilogue

How it finished up. (As-built - schematic attached). One of the patrons at the local blues club heard and desired the amp, so he bought it of me in April, which set me on a path to the next project using this EF86 morph control here: http://www.el34world.com/Forum/index.php?topic=11813.0