Dual Lite JTM45 worth a try?

[Jennings]

I dont have any NFB on this build...so no wire, resistor etc.

[labb]

I would check for a cold solder joint in the wiring around the PI..Maybe the grid of the PI tube. Also the grounds of the PI stage.

[Jennings]

Good suggestion...I did reflow one connection I was suspicious of where the cut control meets the 220k resistors to ground, but I'll go over the whole stage.  I'm also wondering whether there's an issue in the power stage feeding the PI or V3 plate, and the next link down to the triode feeding the PI. 

On a separate note, I'm debating later on whether to dispense with the preamp 220k mixer resistors and use the spare tride in V3 to make each preamp have it's own dedicated stage prior to the PI, then feeding each side of the PI separately.

[Jennings]

Brief update - Part way through separating my ground bus into more of a star ground configuration.  My original bus design worked fine with other builds, but separating the control array onto its own ground line seem to have helped a bit on this one.  When I get more time I'll continue so that the power line, each preamp channel and the main amp circuit all have separate lines.  I've also done away with the mixer resistors and used one half of V3 in each preamp channel, putting them into either side of the PI separately now.  The main issue is the channel interplay still...the slightest nudge of one of the volumes and the amp is very loud instantly.  Turning them up more leads to feedback and squeals.

[Jennings]

After some busy months and a family bereavement I'm aiming to get back to finalising and finishing my KT66 Dual Lite build...first step is to confirm the schematic.  I'd love it if the more experienced could kindly cast a knowledgeable eye over the attached schematic for any comments, corrections, ideas or recommendations.  All gratefully received!

[sluckey]

Schematic looks fine. I do wonder if you need the extra gain stage between the volume controls and the PI. I'd try pulling that tube and connect the volume wipers directly to the PI input caps. Should probably eliminate the squeals. How does that sound?

[DummyLoad]

the EF86 and paralleled 12AX7 share the same power node: speaking from much experience building similar front ends, you really want a seperate supply node for each of those stages and a 100nF HF shunt on the EF86 B+ supply.

grounding is critical. ALL preamp part ground connections should follow one bus to a single ground point at the input jack, IF you are using a switchcraft jack, if you're using a cliff jack, take the jack's ground lug to V1 and V2 bypass caps ground and V1 and V2 power supply caps grounds then wire those to a ground lug near the input jack.

don't bypass gain stage that drives the CF - if you want more gain - bootstrap the CF as shown in the attached. you don't have a gain sucking tone stack after the CF, so you're slamming the LTPI VERY hard. consider divider with a fixed R and pot to tame that - again, see attached. R15 can be 1M or more or make VR4 smaller.

use coax from VR1 and VR2 to the 220K mixing Rs and mount the mixing Rs as close as possible to V3 pin 7 - use a terminal strip.



--pete

[Jennings]

Thanks guys...much appreciated. And that Dual Lite 45 schematic is a great reference for me. I’ll do the minus gain stage check, and then move on to separating the power node for the EF86 and first ECC83. I’ll also review the grounding etc as suggested.


Once I get this running without the hum and volume/squeal issue I’ll then work out whether I want more or less gain form the pre-PI stage. I’m aiming for a ballpark AC30 response as the volume goes up.


I’ll post the result in due course. And a sound clip to help show what the problem sounds like. I’m still slightly suspicious there’s a wiring fault/short that I just can’t trace or see yet...or a dodgy component.

[Jennings]

I’ve split the ECC83 / EF86 power node with a 1.5k resistor and another 8uF cap, plus tided up the grounding as well. While I was there I found I’d accidentally switched the OT and ground connections to the speaker jacks 🤦🏻‍♂️  Can’t believe I’d missed that! Stopped to test the result, and I’ve succeeded in fixing the hum/issues and the amp controls all work fine, tone is good etc, and its no longer clouding the issue.


So now on to the gain...turning up either volume even slightly gets way too much volume too fast...and the amp just starts feeding back...so definitely the extra pre-PI gain stage either wants removing or reigning in. Before I start, I should just whip it out, or maybe try dialling it down by inserting resistors between the gain stage and the PI? Thought I’d ask before I start on it.

[DummyLoad]

suggested mods attached: individual drivers with added complexity -or- plate summing is second.

before you hack, try removing the bypass caps on the LTPI drivers and replace 12AX7 with 12AU7. 


--pete

[sluckey]

I would temporarily bypass the extra gain stage. Just pull the tube, remove the 100K plate resistors, and put a jumper between grid and plate pins directly on the socket. Play it a bit to see if you're happy without the extra gain. If you find you still want more gain put everything back as it is now. Then put a voltage divider just prior to the PI input cap. You will need a coupling cap between the plate and the two resistor voltage divider. You may also need a voltage divider prior to the grid of the extras gain stage.

[Jennings]

Thanks for the guidance...much appreciated as ever...I’ll give that a go 👍


Update: I’ve removed the tube and bypassed the extra gain stage. Quite like the result, and it definitely makes the amp useable now. I’ve only tested it through a bass speaker I had to hand this far, but tomorrow I’ll pop it in it’s 1x12 combo can and try Celestion Century for a bit. Could be a little tame, but it’s hard to be sure on a brief test through a bass heavy cab. Great progress...and thanks for the assistance!

[pdf64]

The standby switch arrangement is ‘hot switching’, not good for good service life from the rectifier tube.

[Jennings]

Crikey...I'd completely over-looked that!  Thanks!  I normally never use standby switches, and only included one here as I had the chassis hole...I'll move it.  Actually I might disconnect it and just leave a dummy switch to avoid the faceplate hole.

[sluckey]

I just noticed that you used a CK 7211 (on-on-on) switch. I would think that when you select both channels the sound would get thin because the signals will be out of phase and do a lot of cancelling. The signals are in phase up to the PI inputs, but the "out of phase" condition is due to the fact that each signal is applied to different inputs of the PI. The LTP PI is really a differential amplifier, meaning it amplifies the "difference" between the two input signals. If the two signals were exactly the same the PI output would subtract those signals, producing zero output at the plates.

BTW, I just used a simple DPDT switch on my Brit Dual Lite for this very reason. So, how does your amp sound when both inputs are selected?

[Jennings]

That makes sense...it does indeed sound thin on the middle setting.  I might consider mixing the channels earlier and feeding into one input of the PI instead as part of considering Pete's mod suggestions.  I might also replace the switch anyway though, as I'm probably fairly unlikely to use the middle setting much really. 

Out of interest, would there be any benefit in me mixing the channels via resistors into one input of the PI, then referencing the other to ground over my current set up (or vice versa)?  Other than saving one half of an ECC83, and thus also needing one voltage divider only rather than dual...thinking of if I do go back to having the extra gain stage in there and a voltage divider as per Pete's schematic mod suggestion.

[sluckey]

Out of interest, would there be any benefit in me mixing the channels via resistors into one input of the PI, then referencing the other to ground over my current set up (or vice versa)?

Absolutely. Doing that keeps everything in phase and will allow you to select both channels. I like Pete's mod3 idea. I would probably experiment with replacing his 220K and 100K voltage divider with a 220K and 250K pot. Might make it easy to dial in just what you want.

[Jennings]

Thanks Steve, much appreciated. I’ll try your and Pete’s suggestions out for size and see where we settle out at.


I like the trim pot idea a lot combined with mod3 if I put the gain stage back in...one daft question though, would I need to add summing resistors in when bringing the plates together, or are they just fine as drawn in mod3?

[sluckey]

would I need to add summing resistors in when bringing the plates together, or are they just fine as drawn in mod3?

No summing resistors needed. The tubes provide the mixing.

[DummyLoad]

if you try mod3 - try with both plate resistors first, and then with one plate resistor.

you could also add switchable bypass to either or both stages individually for more bite from a specific channel.

suspect 1 plate resistor with both triodes unbypassed will sound best overall.


--pete

[Jennings]

Thanks for the help and guidance...that all makes sense, and is much appreciated.  I've removed the standby switch, as I never use them anyway.  I've also tried removing the gain stage, after some tube rolling with ECC82 and 81...then rebuilt the gain stage to Mod3, one plate resistor, with an ECC83, and as predicted it's bob-on what we need...the right level of gain etc.  I was going to insert a trim pot in case I had to tune it, but stopped short as it's behaving just fine as-is and I'm happy with the result.  I stopped short of adding switching for gain levels for now...probably too many options for my simple guitarist's mind!  Haha!!!  Might try some more switching options on future builds though...I'm learning a lot from you folks, and these builds, and this is the first time I've experimented with dedicated gain stage options like this.  One last task for me before I complete this project and build up the combo to completion...there's a return to a little bit of channel inter-play that I just need to stop.  Turn up the volume on the channel that isn't selected, and you get some noise.  Not like before...this time it's much less and lower in volume.  Sounds like a grounding thing to me, so I'll check through my wiring and review where I've grounded the voltage divider 100k resistor on the preamp ground rail. 

[Jennings]

Had a quick run through with the amp last night...all looks in order, and it sounds great in its 1x12 birch ply cab.  This one I've put with a Celestion Century Vintage, mostly for weight saving.  Very quiet at idle, and very low hiss/hum in operation too, so I'm happy the basic circuit is in order.  Just the one niggle remains.  Select the 18W channel and turn the Vox volume up, and you get a bit of hum when you reach 12 o'clock on the volume knob.  Select the Vox channel and turn the 18W volume up, and you get a sort of out-of-phase (hollow, fairly thin) sound of the guitar...almost like it's a deliberate feature, and it's as loud as a deliberate channel on the 18W side.  No crackles, no buzz/hum associated...it's pretty cool actually, like when you have two pickups out of phase on a guitar...but it's clearly not supposed to be happening.  So long as you keep the volume at 0 on the unselected channel this issue doesn't cause a problem when using the amp.  I've tried running the control pot gounds into their respective tube cathode grounds with the selector switch ground joining the preamp bus where it leaves the board and chassis-gounds at the input.  I've tried running the controls on their own bus (Vox to switch to 18W to input jack to main gnd point).  I've tried running a ground from each control side to the switch, and then from switch to the jack ground (current set up).  None of these approaches were noisy, but none seem to impact the channel issue either.  I can live with it for now and use the amp, but it obviously niggles me a bit.  What do folks think?! 

[DummyLoad]

the second gain stage (the summing amp) is picking up noise from the PS of the inactive channels input tube when the volume of the inactive channel is turned up. IOW, noise is leaking through the anode resistor of the input EF86 or input 12AX7, whichever one is the inactive device and being amplified by the inactive channels summing amp. this is one of the drawbacks of this scheme. to effectively deal with problem, you need to short the input of the summing amp of the channel that's inactive. splitting the PS node of the input tubes to separate PS taps MAY help with this and cross-talk, e.g., each channel input tube and summing amps get their own respective PS tap.

for this scheme you have the solution - turn the volume down of the unused channel.

if you decide to walk down the path of electronic switching, then ground the input of the preamp and input of the summing amp of the inactive channel.



--pete

[sluckey]

I think this will solve your problem...

Remove the switch from the input, wiring the input jack to both 33K resistors. Use a SPDT (center off) switch, outer lugs connected to the wiper of each volume control and center lug connected to ground. Should fit neatly on the control panel.

[Jennings]

Thanks again gents...not only for the build help and patience, but also the switch suggestions.  And thanks to all the folks from earlier on with various suggestions, hints and input.  I'll probably get some time over Christmas to open her up and try Steve's switch suggestion.  I'll order one to have a play with, and let you know what happens.  In the mean time I'll enjoy her in her current configuration...it's a compact, open back, birch ply 1x12" combo cab (a couple of inches deeper to accommodate the KT66 with speaker clearance), Celestion Century Vintage speaker (60W).  Valves are two TAD KT66, a Mullard GZ34, Telefunken EF86, and the ECC83s are all Electro Harmonix.  I might.  This build and combo combination gives a bigger, fuller amp than my previous 2xEL84 build in a normal depth cab with alnico speaker.  I like the contrast, and I'm really pleased with the tone.  I've got some practices and gigs in the New Year to try it out in anger.  Here are a few pictures of the completed build:

[Jennings]

And here is a revised schematic with the final (for now!) revisions, a couple of slight value corrections here and there for trueness to the build, plus I've indicated the grounding groups.