Hoffman Amplifiers Tube Amplifier Forum
Amp Stuff => Tube Amp Building - Tweaks - Repairs => Topic started by: plumcrazyfx on March 17, 2018, 09:54:25 am
-
Looking for input on a design I've been working on - it's a modified 5C1 with an EL34 and an EF86 parallel input with the 6SJ7. This started with wanting to have a really simple build that I could have parallel pentode input - figured a lot of designs allow you to run 12AX7s parallel, so give it a try. So, I have a working prototype with leftover parts that I have sitting around here (smaller transformers) but once I have it tweaked I plan to go with a larger power and output transformer and a final build with new parts. The voltages on the schematic are generally right - I've been doing a lot of tweaking so I can't say that they are all the most recent - but they're close.
This started as 2 6SJ7s parallel with one side with slightly larger values. Didn't really move me so I converted one to cathode instead of grid leak bias. Still not quite what I wanted so I used an adaptor to put an EF86 on one side. After that I tweaked values to try and get the voltages into a reasonable range and tweaked some things based on how they sound. Right now the thing has a rootsy grit when all of the way up. The 4 inputs give 4 different levels of gain which works for me - I didn't want a Tone knob because I don't really use them and didn't want to rob the circuit of any gain. I dig the low overdrive as a good pedal platform. The thing is a raving gain monster with something like a Rangemaster driving it but doesn't cut out as I've heard some say pentode can do.
Sorry about the upside down scan - I don't have access to my scanner right now to fix it.
-
The control grid (pin 9) of the EF86 must have a resistor to ground. 1M works for me. Also, pins 2 and 7 should be connected to ground.
-
I forgot to draw it on the schematic but I do have the 1megs on the input jacks - does this work?
I missed the shield. They are grounded.
-
I forgot to draw it on the schematic but I do have the 1megs on the input jacks - does this work?
No it doesn't because of that .047µF cap. You don't need that cap. Either replace the cap with a short piece of wire or add a 1MΩ right on pin 9.
-
Got it. What is it doing without that on there? It's working - is it dangerous, bad for the tube, or just not good design practice?
-
It ain't working right. Tube bias is determined by the dc voltage between the control grid and the cathode. But if the grid has no dc path to ground and back to the cathode, there can be no bias voltage established. Tubes can behave in a number of odd ways with no defined bias.
Try it. Just short out the cap (my choice) or add a 1MΩ directly to the grid.
-
Will do. Will report back.
-
Will that shift around all of my other values on the EF86?
-
Will that shift around all of my other values on the EF86?
Shouldn't. If you want to see some proven circuit values look at this schematic...
http://sluckeyamps.com/VAC15/Vox_AC15.pdf
-
Pulled out the .047 so now the control grid sees 1meg + 68K. The voltage on the plate jumped to 155V and the screen jumped to 119V. Sounds basically the same. All other voltages in the amp stayed the same.
-
I looked at both that and Dr Z stuff for initial adjustments for the EF86 but kept some of the stuff from the 5C1. I kind of want to build it for the 6SJ7 and continue using the adaptor for the EF86 or 6AU6 - does wiring it without the .047uf still work with other pentodes? I suppose the grid bias already has that with the 5meg.
-
I kind of want to build it for the 6SJ7 and continue using the adaptor for the EF86 or 6AU6 - does wiring it without the .047uf still work with other pentodes? I suppose the grid bias already has that with the 5meg.
The cap has nothing to do with which tube you use, pentode or triode, makes no difference. The cap is used in conjunction with that big 5M grid resistor and the fact that the cathode is tied to ground. That defines a "grid leak bias" circuit. If you want the EF86 to operate as a grid leak circuit then put the cap back, put a 5M resistor from grid to ground and also connect the cathode directly to ground.
Right now, your EF86 is operating as a cathode biased circuit. You can use the 6SJ7 in a cathode biased circuit simply by changing that 5M to a 1M and putting a resistor/cap on the cathode.
The old grid leak circuit went out of style during the mid '50s and was replaced with the superior cathode biased circuit. Google these type circuits for more info than you probably want.
Here's another cathode biased pentode with slightly different tube and components that I recently built...
http://sluckeyamps.com/smoky/smoky.pdf
-
I'm fine with having the 2 styles of biasing in the same amp. I plan on mainly using EF86 or 6AU6 (which also sounds really good) on the cathode biased side but wanted to retain the ability to use the 6SJ7 cathode biased if I wanted to. I might be doing a tweak on that side again to slightly lower the voltages on the cathode biased side. Looks like I could be back to the 1m/220K combo from the AC15.
-
The connection to the volume pot is like in the schematic or you used mix resistors to connect the two channels or a SPDT switch ?
Franco
-
No switch. Just connected it in there.
-
I was convinced that the properly connection of two channel must be performed using mix resistors or a switch not just connect it
abruptely in parallel to the following stage
that is totally new to me
BTW, you talk about paralleled tubes but your schematic show the output connected in parallel but not the imput
in the real circuit did you give a try to a paralleled input and output ?
(https://i.imgur.com/WH4BeL9.jpg)
Franco
-
It requires a jumper between inputs to run the two preamp tube parallel. I don't know if connecting both the 6SJ7 and the EF86 to the same point on the volume is the best way of achieving this idea. I was thinking of the 5E3 without the 2 volumes when I threw it together. It seems to work - when you jumper there is more gain and a much fatter sound which was the intention. If someone has an idea about a better way, I'm open. That's why I posted it here - to head off something I don't want in the final amp. I do dig the sound as it is and the choice of the low or high input for each tube and then the ways of jumpering it give me a lot to work with.
-
I don't know if connecting both the 6SJ7 and the EF86 to the same point on the volume is the best way of achieving this idea. I was thinking of the 5E3 without the 2 volumes when I threw it together.
A better way would be to put a 220K resistor in series with each coupling cap and connect each 220K together at the top of the volume control.
The 5E3 circuit is unique. It has the volume pots wired "backwards" so the pots actually act as mixing resistors unless you crank both pots full up.
-
To link the inputs you have more than one chance
Hiwatt
(https://i.imgur.com/2dsTJmd.gif)
Marshall
(https://i.imgur.com/1zM5A2E.gif)
VOX AC30CC2
(https://i.imgur.com/RrOE9y1.jpg)
Simplified VOX diagram
(https://i.imgur.com/HGeP6nO.jpg)
Franco
-
I like the linked Hiwatt version but does that really only save me one jack in terms of build complexity? I also seem to notice some difference between the way I link them either into the hi for the 6SJ7 out of the lo and linked into the hi of the EF86 or if I do it reverse - might just be perception?
Does the value of the mixing resistor matter or just that they are the same? I don't want to rob the circuit of any gain at this point since dig the grit I get (it's just enough - sparkly grit with single coils - a bit more with humbuckers, fatter with linked channels. Could I do 100K, 47K, 22K? What it the purpose of the mixing resistors at that point?
-
Click here (https://www.google.com/search?q=mixing+resistors&sourceid=ie7&rls=com.microsoft:en-US:IE-Address&ie=&oe=&rlz=1I7GGLL_en&gws_rd=ssl)for the ultimate authority on mixing resistors...
-
Reading here
http://www.valvewizard.co.uk/resistive.html (http://www.valvewizard.co.uk/resistive.html)
.................... Later stages in the amp will usually require larger mixing resistors to ensure that even if one channel is grounded (gain control turned down) the input impedance seen by the other channel is not too low (roughly equal to the two mixing resistors in series). Values of 100k to 470k are typical, but beware of unwanted high roll-off due to the input capacitance of the following stage (see the Miller effect). In this case the gain controls are providing the necessary grid-leak resistance. The mixer resistors must come after the potentiometers or it would be possible to short either channel to ground by turning one or other of the potentiomers to zero.
(http://www.valvewizard.co.uk/Resistive3.jpg)
seems that in your circuit (as it is now) is not strictly indispensable to use it
Franco
-
just curious...
1) what's up with the filament wiring and connecting one side of filament secondary to 100R to cathode of EL34?
2) why do you need a 2W 1.5K grid stop for G1 of EL34? control grid will have vaporized long before you could flow enough current through 2W resistor.
3) maximum grid (G1) ckt. resistance for EL34 is 500K. use a 500K pot? parallel 510K R with pot wiper. phillips states .7M for class A/AB mode or .5M in class B; mullard states 500K max.
--pete
-
These are the reasons I posted it here - to sort these issues out before I finalize it. Both 100 ohms go to the cathode to lift the heaters. Might put a humdinger in.
I can lower the grid stop if it is meant to act as circuit protection.
I can lower the volume to a 500K if that it necessary. Will that still make it suitable for changing to a KT66?
-
thanks for the clarification - your sketch of filament string is confusing.
usually 1.5k 1/2W is more than adequate for grid stopper.
for KT66 500K - if in doubt, check specs of KT66 for maximum recommended g1 circuit resistance. attached to this reply - hint look under section "general".
as pointed out by K and others - individual level controls and summing resistors would be nice.
--pete
-
some documentation to help. please ignore classistone OT reference this schematic was transposed from another schematic: will fix later.
--pete
-
Nice Pete :thumbsup:
Franco
-
Pete, you may want to move R2 to the other side of C1.
-
Dang guys. Brilliant. The level of knowledge here is unbelievable.
-
Pete, you may want to move R2 to the other side of C1.
opps! :icon_biggrin:
--pete
-
Nice Pete :thumbsup:
Franco
thanks franco!
--pete
-
fixed OT, notes edits., date, etc., and error as pointed out by sluckey. cleared attachments in previous reply - please use these drafts.
fill in the rest of the telemetry if you wish and i'll update the draft.
--pete
EDIT: R12 is redundant. eliminate it.
-
I will try to make some changes and get back next weekend.
-
deleted R12. fixed ref-des. R1.0a last rev.
--pete
-
Going to do some work on it this weekend but wanted to get a jump on some questions/comments.
Is there a benefit to going with 500k pots with the EL34?
R12 was deleted. I was using a 1.5k directly at the tube as a grid resistor. Is this not needed with the mixing resistors and 2 controls?
The voltages on the EF86 jumped when I removed the unnecessary cap. R6 is now 180k and R9 is now 1meg. These brought plate and screen voltages down into the 100v range. (Will get new voltagrs on everything after tweaks).
Any reason most cap values changed to .047uf?
Thanks.
-
Going to do some work on it this weekend but wanted to get a jump on some questions/comments.
Is there a benefit to going with 500k pots with the EL34?
R12 was deleted. I was using a 1.5k directly at the tube as a grid resistor. Is this not needed with the mixing resistors and 2 controls?
The voltages on the EF86 jumped when I removed the unnecessary cap. R6 is now 180k and R9 is now 1meg. These brought plate and screen voltages down into the 100v range. (Will get new voltagrs on everything after tweaks).
Any reason most cap values changed to .047uf?
Thanks.
R12 NOT needed.
500K pots are fine.
.047uF are personal preference. use whatever value suits your ears. some prefer .022uF. same with cathode bypass caps - experiment. clip leads help here: you choose the value before soldering in the part.
--Pete
-
Thanks. Will post voltages.
-
Thinking about putting in a humdinger with the mods this morning. I have an old Honeywell 500ohm 2W pot. Am I okay still lifting the heaters? I was just going to take the middle lug to the cathode/cathode resistor junction of the EL34 and the ends to the heaters. I have one on my Bassman 20 and like having it but I don't know if I've seen a discussion of using one with lifting the heaters.
-
If you mean if is possible to have an humdinger and elevated heaters at the same time
Yes, you can
Franco
-
That was the question and thanks.
-
Made the changes and it sound good.
The voltages are:
6sj7 - 1, 3 & 5 - grounded, 2 & 7 - fil (24v), 6 - 22v, 8 - 122v I don't think my meter is very good - I can't read any voltage on pins 4 & 5
ef86 - 1 - 104v, 2 - 0, 3 & 8 - 2.3v, 4 & 5 - fil (24v), 6 - 111v, 7 - 0, not reading anything on 9
el34 - 1 & 8 tied - 24v, 2 & 7 - fil (24v), 3 - 391v, 4 - 358v, 6 - 366v, not reading anything on 5
-
If you decide to buy a new meter, better a True RMS model
if you want only a cheap one (among true RMS)
https://www.banggood.com/it/ANENG-AN8009-True-RMS-NCV-Digital-Multimeter-9999-Counts-Backlight-AC-DC-Current-Voltage-Tester-p-1216900.html?rmmds=search&cur_warehouse=CN (https://www.banggood.com/it/ANENG-AN8009-True-RMS-NCV-Digital-Multimeter-9999-Counts-Backlight-AC-DC-Current-Voltage-Tester-p-1216900.html?rmmds=search&cur_warehouse=CN)
https://www.banggood.com/it/ANENG-AN8008-True-RMS-Digital-Multimeter-AC-DC-Current-Voltage-Resistance-Frequency-Capacitance-Test-p-1157985.html?rmmds=search&cur_warehouse=CN (https://www.banggood.com/it/ANENG-AN8008-True-RMS-Digital-Multimeter-AC-DC-Current-Voltage-Resistance-Frequency-Capacitance-Test-p-1157985.html?rmmds=search&cur_warehouse=CN)
Franco
-
6sj7 - 1, 3 & 5 - grounded, 2 & 7 - fil (24v), 6 - 22v, 8 - 122v I don't think my meter is very good - I can't read any voltage on pins 4 & 5
brain fart, right? pin 5 is GROUNDED! lol! ground is 0V... :icon_biggrin:
measure across pins 2 & 7 with meter set to AC for filament voltage - you're reading the DC voltage of the EL34 cathode that's tied to the heater artificial center tap.
don't bother trying to read voltages of the control grids (g1) your meter loads it too heavily for an accurate reading. those are: pin 4 of 6SJ7; pin 9 of EF86; pin 5 of EL34.
--pete
-
if you'd be so kind to take some more measurements to finish updating schema voltages: with the amp and warmed up (~10 mins), measure V4 between pins 4 & 6 with meter set to ACV, also measure V4 pin 8 to ground with meter set to DCV; finally, measure V4 between pins 2 & 8 with meter set to ACV.
thanks,
--pete
-
Will do.
The thing sounds great. Just the right amount of grit. I've been tweaking it with and old Bassmate cab with the 15" RSC speaker with the back off and it is a good match. I've been testing it with a JJ E34L and switched to a Sino EL34 which was louder while flabbier in the bass. Tried a Sylvania 5881 and it was too clean - the KT66 sits in-between the clean an grittier EL34 and is nice too. I still have the 4 inputs but I may try the Hiwatt 3 input to give each channel or the jumpered. Right now I have the high or low inuts, those input change slightly when the unused volume is adjusted and then there is the ability to blend the two channels when jumpered. The humdinger and 47ufs for the first 2 caps makes it silent with a humbucker or hum cancelling setting despite the proto being a mess of wires.
Is all of that filtering still okay with the GZ34?
-
V4 - 2 to 8 - 5.39v
4 to 6 - 661v
8 to gr - 421
V3 - 2 to 7 - stupid Harbor Freight meter this one jumps all over the place - maybe a new battery.
-
yes. 50uF filters OK with GZ34. datasheets state 60uF max. in most conditions and you're pulling less than 1/2 max. rated DC steady state output current.
--pete
-
updated telemetry.
--pete
-
dually achieved :laugh:
-
If I want to do the Hiwatt jack switching, which jacks do I need to use - 14B or 114BX?