ODS 124 - revisited for the 21C

[Stephen1966]

Thanks Steven I appreciate that!

So I will give you my take and you can do with it what you will..

Tony

Thanks Tony. Likewise...

I see it as some kind of balance... Personally, I think the part selection is crucial as you say, especially where the signal path is concerned. Hence, I have a bag of 6PS caps sitting on my desk waiting to tested and measured. As you mentioned, it starts to look pretty straightforward as you break it down into its constituent parts but then you put all these gain stages together and it becomes another kind of beast altogether. Another thing I will be doing is using carbon film resistors... some people don't seem to think very highly of them but when Dumble used them, it wasn't because he couldn't get hold of metal film. I think its healthy to assume that everything that is there in the 124, is there for a reason. I may not understand those reasons but that is no excuse to go ripping it apart with ill considered, baked-in changes. The fundamental circuit is to be maintained as much as possible and any mods, should be switchable so that with them out of the circuit, what remains is the essential 124 "neurology," as unabridged and unadulterated a it can be. I agree, it really makes sense to stick with the parts list as much as possible. The power section has to be overhauled though, and Martin's precision power supply seems a no-brainer. As a sidenote, I recently found some Ruby caps and I am going to use these in my Tweedle Dee in place of the F&T's that are in there at the moment. Maybe filter caps do have an effect on tone, but if so it is probably quite subtle, if noticeable at all. I will be getting back into that project and sorting it out before I order the power supply caps for 124 so it's a chance to test for any differences that may occur.

I do have one question which has been bugging the hell out of me, perhaps you could answer this one...

In the photos of the 124 which I posted earlier today, it's clear that D used a different type of resistors for the first triodes' cathodes of both the clean and overdrive channels. Gil's sketch [Edit: the sketch was made by Bill, if I understand correctly] marks all four as 1.5k but there is a distinct difference between the fatter, larger resistors on the input side, second triode ( pin 8 ) and the smaller RN65 form of the white resistors on the output side ( pin3 ). The 123 schematic you posted, also has different values for the cathodes of single tubes. For the moment, I don't fully understand why that might be necessary, though if I recall correctly, it has something to do with the Miller capacitance of the first and second triodes. To me, the value of the resistors looks quite important. Could it be that the input cathode resistor has a higher rated Wattage? Or is a different type, ceramic wirewound or so on? It's a shame we don't get to see it in its naked state without all the (expletive) goop, but even that was there for a reason. If anyone has any ideas there, it would be good to hear them.

Yeah, so, someone once told me that as more mods are added there is more chance the endeavor will go south. I think it may have been you, in which case, thank you, I agree. The Tweedle Dee entailed a fair bit of reading up before and after I put the parts on the bench to be assembled, but the 124 is next level stuff. I don't want to be somebody's kit-builder. I really want to understand this Voodoo Magic... [English is such a paltry language - "I want a chocolate amp with a syrupy hint of mojo. Hell! make that a double!"]

Stephen

[Stephen1966]

Correction, I called the overdrive trimmer a variable resistor earlier - it is if there is no connection as indicated in Bill's sketch - but the layouts, schematics ad infinitum show it really is as Gill said, a voltage divider.

Geez! It's been a long day. Time to turn off the internet

[pompeiisneaks]

The first input jack seems to have no ground reference for the 1M resistor... You might just need a ground connection at the bottom of the Switchcraft 12A jack to get your ground you intended there.

~Phil

Thanks for that Phil, it's a simple job to restore the GND trace and, for the sake of completeness, I should do it anyway (the FET jack has one). Forgive me if I'm wrong but aren't the sleeves of all Switchcraft jacks connected through their bushing to GND anyway? I will put it back though, because the next builder who comes along might decide to use an isolated jack, and then as you say, it needs the GND.

That's definitely an assumption that not all are going to use. Some might prefer cliff style jacks that are always isolated etc. besides, the ground of the jack to the chassis even if assumed should still be on a schematic.

Also it's always a smart practice to wire a lead from the ground lug to a ground buss anyway because corrosion or vibration over time can make jacks wiggle loose and lose their ground.

~Phil

[martin manning]

Edit: Notice Martin has the OD entrance resistor marked 120K is wrong it's a 220K on the boards

Final version of that board is here, second to last post in the thread: https://ampgarage.com/forum/viewtopic.php?f=4&t=6020 The power supply boards are there too. If I were building today I would use the radial cap versions.

[talbany]

In the photos of the 124 which I posted earlier today, it's clear that D used a different type of resistors for the first triodes' cathodes of both the clean and overdrive channels. Gil's sketch [Edit: the sketch was made by Bill, if I understand correctly] marks all four as 1.5k but there is a distinct difference between the fatter, larger resistors on the input side, second triode ( pin 8 ) and the smaller RN65 form of the white resistors on the output side ( pin3 ). The 123 schematic you posted, also has different values for the cathodes of single tubes. For the moment, I don't fully understand why that might be necessary, though if I recall correctly, it has something to do with the Miller capacitance of the first and second triodes. To me, the value of the resistors looks quite important. Could it be that the input cathode resistor has a higher rated Wattage? Or is a different type, ceramic wirewound or so on? It's a shame we don't get to see it in its naked state without all the (expletive) goop, but even that was there for a reason. If anyone has any ideas there, it would be good to hear them.

The value of the bypass caps goes? I've heard them range from 4.7uF all the way to 50uF..Much of this depends on the tonestack and plate loads used with the different value voltage dividers (Pot's) throughout the amp.So if you want something to experiment around with , there you go.
As far as the cathode resistor types go IMO these are less critical for tone shaping purposes. I would pay larger attention to things like Slope/Plate's/OD feeders/Grids and to some extent Bypass cap types
IMO Dumble preferred the Precision metal films there mostly for their stability, temp. coefficient and tight tolerances. In most amps i have seen Dumble liked to keep each side of the triodes symmetrical and balanced.

Tony

[Charlie Wilson]

The cathode resistors in #124 are Mepco Electra(MFC62) and IRC plastic molded metal film. They both are 1.5k. Both those resistors sound a bit different. The IRC is smooth and kind of flat sounding and the Mepco is a bit rawer sounding on the top end. I believe he combined both to add a bit texture or edge to the top and also a bit more 3D sounding. 124 is a low plate with 100k plates and 1.5k cathode resistors, 123 is a high plate classic with 220k/150k plates and 3.3k/2.2k cathode resistors. Also, those cathode resistors see very little voltage, I agree with Tony but I believe the #1 reason Dumble uses any particular component is for what they sound like. All of the carbon film resistors on the board except for the NFB resistor are less than 1/2w. They are LCA 0411 Sprague Q-Line rated at .4W.
CW

[talbany]

The cathode resistors in #124 are Mepco Electra(RN62) and IRC plastic molded metal film. They both are 1.5k. Both those resistors sound a bit different. The IRC is smooth and kind of flat sounding and the Mepco is a bit rawer sounding on the top end. I believe he combined both to add a bit texture or edge to the top and also a bit more 3D sounding. 124 is a low plate with 100k plates and 1.5k cathode resistors, 123 is a high plate classic with 220k/150k plates and 3.3k/2.2k cathode resistors. Also, those cathode resistors see very little voltage, I agree with Tony but I believe the #1 reason Dumble uses any particular component is for what they sound like. All of the carbon film resistors on the board except for the NFB resistor are less than 1/2w. They are LCA 0411 Sprague Q-Line rated at .4W.
CW

I'll Bet!
I can't speak for the IRC's (impossible to find the MEB's) however here are the Plastic molded Metal Film Dale RN75-B's I put them in to see if they would help smooth out the OD and tighten the bass any?..That they do!

Tony

[Stephen1966]

Phil, Martin, Tony, Charlie. Thank you. Sincerely. That is a lot to digest but your observations answered a number of questions I have been mulling over, and poses a few more. There is a lot to take in here though so I'll start with the most straightforward.

Phil. Thank you. I've been reading schematics for a while now but it is only recently I learned how to draw them. A great learning tool if nothing else, but it means I am not totally familiar with the conventions we should be using. However, your point about the mechanical properties of jacks is well taken. For these reasons, and more, it makes sense to mark the ground on the schematic - a seperate one for each ground even if they are connected along the way in a variety of ways I've seen from photos. It just makes sense logically, to include the GND traces if only because it is easier to adapt something that is there depending on the type and configuration of jack choices, than it is to work with implied connections.. however correct they may be.

I'm working on a revised version of the schematic which points the jack grounds to GND point 1 on the map and I will post this a little later, when I've worked through the ideas being presented here.

P.S. which is correct BUS BAR or BUSS BAR?

[Stephen1966]

Edit: Notice Martin has the OD entrance resistor marked 120K is wrong it's a 220K on the boards

Final version of that board is here, second to last post in the thread: https://ampgarage.com/forum/viewtopic.php?f=4&t=6020 The power supply boards are there too. If I were building today I would use the radial cap versions.

Martin, thank you. This is a post I hadn't come across before and it is very interesting in the way it documents the process of drawing up the layouts and where you found your sources. It also inlcudes the FET board PCB from ic-racer. Your eyelet boards are good as well, but I could easily produce the PCB if it checks out. The main boards however, are just great. I would have to work them through Illustrator as I don't have Visio and match actual component sizes with the eyelets (vice versa actually) but this is a great job and saves me another headache. Thanks again.

Good health!

Stephen

[pompeiisneaks]

...

P.S. which is correct BUS BAR or BUSS BAR?

I think it's BUS, but I always had BUSS fuses around when I was young and I think that's getting confused in my head

~Phil

[ChopSauce]

I also think it's bus, like in: "all the electrons take the same bus to go down on ground"...

(yet another) awesome post, by the way!

[Stephen1966]

I also think it's bus, like in: "all the electrons take the same bus to go down on ground"...

(yet another) awesome post, by the way!

Thank you ChopSauce, and Phil. I agree. I'll stick with Bus. I'm an Englishman after all

[Stephen1966]

The cathode resistors in #124 are Mepco Electra(RN62) and IRC plastic molded metal film. They both are 1.5k. Both those resistors sound a bit different. The IRC is smooth and kind of flat sounding and the Mepco is a bit rawer sounding on the top end. I believe he combined both to add a bit texture or edge to the top and also a bit more 3D sounding. 124 is a low plate with 100k plates and 1.5k cathode resistors, 123 is a high plate classic with 220k/150k plates and 3.3k/2.2k cathode resistors. Also, those cathode resistors see very little voltage, I agree with Tony but I believe the #1 reason Dumble uses any particular component is for what they sound like. All of the carbon film resistors on the board except for the NFB resistor are less than 1/2w. They are LCA 0411 Sprague Q-Line rated at .4W.
CW

Thank you Charlie. What is interesting to me here is that he combined both the Mepco Electra and IRC on opposite sides of the tube. The larger resistors (the IRC if I understand it correctly) are going to ground from the second cathode ( pin 8 ) and the Mepco Electra are going to ground from the first cathode, the output side of the tube, ( pin 3 ). I too, am going with the low-plate recipe of the 124 though I can see it wouldn't take much to turn it into a high-plate like the 183. Again, if I understand it correctly, most of the amps after 124 went to a high-plate configuration. I do hear good things about the dynamic range of the 124 though so it is my starting point. I'll live with it for a while and perhaps even just build another in a high-plate configuration later on.

The thing I find so valuable about this post is just how knowledgeable you are when it comes to identifying and talking about these component choices. The Draloric brand (AKA Q-line) resistors consists solely of a thin film inventory these days but Draloric, and Sprague appear to have been bought out by Vishay. For someone like me, entering the field so recently and with so many options on the market you can probably imagine just how daunting and formidable it can seem. IRC resistors are still available in ebay-like stores (very dodgy) but there is nothing like the Mepco Electra values I can find. Given your vast experience, do you, or indeed does anybody, have any thoughts about what the modern and easily available equivalents might be? Components that have similar sonic properties.

Good health!

Stephen

[Stephen1966]

This post carries on from the information Charlie gave us about the component types and I want to pick up with where he left off with info that most of the Q-line resistors were around 1/2W. This question is open for everyone, not just Charlie though and it gives me a chance to talk about another aspect of D's work I've been thinking about for a while.

I have a theory I would like to pass by you. It goes like this... I'm not so surprised that the majority of CF resistors are just around 1/2W. Most Fender designs use these as stock values but it does bring up another aspect, a trademark if you like, of D's design. Goop. When I look at most people's designs on here and listen to what they say, they say 1W is probably the norm. Is it possible that the higher value ratings lead to a greater component mass and thus a better handling of the temperature coefficient? Most people here, do not seem to goop their circuits.

One reason why they don't might be because there seems something frivolous or petty about hiding the circuits from techs who come after or about trying to hide something. There may be some truth to that with regard to Alexander. I know lots of human beings and artists who are, sad to say, quite flawed individuals but they still do brilliant things. But I don't buy it. D is too deliberate, too conscious of every small detail it seems, and I just don't buy it was the only reason he gooped his circuits. I have a theory, only a theory mind you, that the silicon is there, epecially in the more sensitive early stages of the amp as a means of thermal control. Silicon is a very poor conductor of heat. Those silicon hand mitts we use are no more than 1/16" thick but allow us to handle objects of hundreds of degrees coming from the oven. The inside of a chassis, with it's host of valves is a very hot environment and subject to extreme temperatures from a cold start up to fully opened up and operating.

It's clear from as far back as the Winterland #003 D was thinking about thermal control.

1d_3.jpg

Silicon has a number of useful properties which can be exploited. It's a terrific thermal insulator, the only heat the components experience is the heat they produce themselves and stacking resistors next to capacitors can be done to quite close tolerances without the heat of one affecting the drift of the other. It also has adhesive properties so good that once embedded in silicon, it's a job to remove it. We see a dab of silicon between caps on the Rock/Jazz switch. There's no attempt here to hide the identity of the components.

ODS_124081.jpg

In other areas of the amp too, Silicon is used as as adhesive. The larger caps are secured this way, relieving the strain on their leads when the amp is moved.

Silicon also has the elasticity to absorb physical shocks, impacts and oscillation which might contribute to microphonic artifacts. The Z5U and X7R ceramic types seem especially susceptible to these but also poor temperature stability (See Aiken Amps https://www.aikenamps.com/index.php/technical-q-a) and yet D used them, even when he had access to the COG dielectric types which have much lower harmonic distortion and stability.

Silicone is also a waterproofing agent. It's airtight in fact and as Aiken pointed out, the cracking of the epoxy seal around the PP caps can lead to moisture getting in and a resulting drift. Silicone solves that problem in one fell goop

And then, of course, a happy accident perhaps, if you use a coloured silicon you can hide your work and protect your "intellectual property". If so, it begs the question, why didn't he also goop the PI, the FET board, the switching or other elements of his work that are characteristically his own. [Edit: okay the PI is not what you could call D's design but it is in the later stages of the amp where signal strength mitigates any stray background noise.]

Maybe there is something to be said for using 1/2W resistors. For running components closer to the extremes of their operating range for ripple currents but then for mechanically heat-sinking them so that they remain at a stable temperature coefficient.

What do you think?

[drew]

P.S. which is correct BUS BAR or BUSS BAR?