Yes !
I use it with a B1 right now. I got a BA3 amp, I'm going to use a separated box for the FE board so that I can switch between preamp...
I also got parts to make an IronPre, a B1K, a Wayne 2018 Linestage, a Pearl 2, a Pass AB100... more fun is yet to come !
I use it with a B1 right now. I got a BA3 amp, I'm going to use a separated box for the FE board so that I can switch between preamp...
I also got parts to make an IronPre, a B1K, a Wayne 2018 Linestage, a Pearl 2, a Pass AB100... more fun is yet to come !
So... reporting on the sonic impact of the additional bypass cap... to the big existing one (10 000uF), which is already bypassed by a PP 0.33uF one.
One thing is clear: whatever the impact, it is quite marginal / very small to our ears. So one might ponder if it is worth doing. On the other hand, it is easy to implement and cost is extremly low.
Gilles' findings were "very subtle difference, on the verge of what is audible to my ears, if ever a difference. After some long comparisons, I believe the bass goes deeper at perhaps some cost of impulse sharpness, quickness. "Perhaps" on the latter one, so perhaps a small trade-off. So small that I can't tell what is better anyway".
Me, I could just do a before / after comparo once I fitted the caps, and these needed some burn-in. My initial impression was "not a world mover, just slighly more groove / bass building from lower registers but no quicker rising front. On the other hand, the latter didn't seem washed off. Other registers not really affected, or perhaps some marginal additional cleaness, although the latter could be some sharpnedd due to the burn-in of the caps".
We tried now many things around this VFET, which was build with connectors to enable several "quick and reversible tests". Bottom line is none of the mods we did on the amp sections did modify the sound in a significant manner: improvements on OS and FE were either small or non existant. That included mainly several cap bypasses and also the bypass of the entire Front End sections.
The one mod that stands out so far is the addition of the PS caps, which clearly addressed a point we felt the original amp was lacking somewhat: impulsivity and slam. That mod is a must do, the others can be done and won't really harm anything while they can perhaps be somewhat benefical to your ears.
Because of that, and unless someone comes up with a bright new idea, we will leave the amp sections now as they are. Papa did a splendid job which can't really be bettered to our ears (Ok, I should have known from the start LOL). The section that despite several claims seems to benefit from some improvements is the PS. Not by a lot: despite me not liking initialy SMPS, the proposed PS does a VG job. BUT the addition of a couple of caps between the existing SMPS filter and the boards did quite a change everyone should be able to hear. So the very last mod we will try will be around additional SMPS filtering sections.
Papa has already provided a simple SMPS filter with this build, so probably (again) nothing to be gained... but this is DIY, so trying apparently silly things can make some sense 🙂
I hope this helps future tweakers: this amp is really a nice sounding one, and time is probably best spent listening to it then soldering on it...
Claude
One thing is clear: whatever the impact, it is quite marginal / very small to our ears. So one might ponder if it is worth doing. On the other hand, it is easy to implement and cost is extremly low.
Gilles' findings were "very subtle difference, on the verge of what is audible to my ears, if ever a difference. After some long comparisons, I believe the bass goes deeper at perhaps some cost of impulse sharpness, quickness. "Perhaps" on the latter one, so perhaps a small trade-off. So small that I can't tell what is better anyway".
Me, I could just do a before / after comparo once I fitted the caps, and these needed some burn-in. My initial impression was "not a world mover, just slighly more groove / bass building from lower registers but no quicker rising front. On the other hand, the latter didn't seem washed off. Other registers not really affected, or perhaps some marginal additional cleaness, although the latter could be some sharpnedd due to the burn-in of the caps".
We tried now many things around this VFET, which was build with connectors to enable several "quick and reversible tests". Bottom line is none of the mods we did on the amp sections did modify the sound in a significant manner: improvements on OS and FE were either small or non existant. That included mainly several cap bypasses and also the bypass of the entire Front End sections.
The one mod that stands out so far is the addition of the PS caps, which clearly addressed a point we felt the original amp was lacking somewhat: impulsivity and slam. That mod is a must do, the others can be done and won't really harm anything while they can perhaps be somewhat benefical to your ears.
Because of that, and unless someone comes up with a bright new idea, we will leave the amp sections now as they are. Papa did a splendid job which can't really be bettered to our ears (Ok, I should have known from the start LOL). The section that despite several claims seems to benefit from some improvements is the PS. Not by a lot: despite me not liking initialy SMPS, the proposed PS does a VG job. BUT the addition of a couple of caps between the existing SMPS filter and the boards did quite a change everyone should be able to hear. So the very last mod we will try will be around additional SMPS filtering sections.
Papa has already provided a simple SMPS filter with this build, so probably (again) nothing to be gained... but this is DIY, so trying apparently silly things can make some sense 🙂
I hope this helps future tweakers: this amp is really a nice sounding one, and time is probably best spent listening to it then soldering on it...
Claude
Last edited:
The PCH-VFET power supply filter board (from the first lottery of ~ 100 amplifiers) uses a radically different design approach than the NCH-VFET power supply filter board (from the second lottery of ~ 100 amplifiers). Check the two .pdfs written by Nelson Pass to see the schematics.
Whichever filter board you have now, try the other. Bare PCBs aren't available so you'll need to build it up using something like Experimenter's Board, or dead bug on ground plane, or peg-boards like guitar amplifiers use, etc. Since those two pdf documents also contain images of Nelson's PCB layouts, you could clone his PCBs fairly easily. Nelson has done the hard part for you: parts placement and copper routing. Voila you now have PCBs which you can use yourself, and hand out to other VFET experimenters.
Whichever filter board you have now, try the other. Bare PCBs aren't available so you'll need to build it up using something like Experimenter's Board, or dead bug on ground plane, or peg-boards like guitar amplifiers use, etc. Since those two pdf documents also contain images of Nelson's PCB layouts, you could clone his PCBs fairly easily. Nelson has done the hard part for you: parts placement and copper routing. Voila you now have PCBs which you can use yourself, and hand out to other VFET experimenters.
Adding some extra supply filter capacitance to the P-type output stage is a surprisingly easy and effective modification. Of course I have gone a step further by building an external linear PSU, but the simple mod is easy for anyone with the capability of building a VFET amp.
I also heartily endorse building one of Mark's new front end boards. Both the Bulwark and Dreadnought designs have been highly entertaining. I expect the others would be as well.
I also heartily endorse building one of Mark's new front end boards. Both the Bulwark and Dreadnought designs have been highly entertaining. I expect the others would be as well.
Last edited:
Thanks for your kind post, Mark.
That is indeed also a potential solution, something I have considered,comething I could still easily build on an experimenter's board...
Apart from the need to probably tame things at start up (many uF) - as NP did with this filter board- the N-VFET is using a RCRC filter (so 2nd order overall with here Fc = 53Hz). It adds indeed many uF that can also help addressing other concerns while at it.
It would require proper modeling to find out how the various Zin and Zout at the filter's interface interact with the filter itself to design a dedicated filter. And the ESRs of the caps - as the ones of the caps I have already added at the OS board PS entry - have also a clear impact.
But that could take me as much time as thinking about it "roughly", as I did in my post 2 months ago, and use the spare time to... just give it a try.
Therefore, at this stage, to evaluate the sonic impact, I will as next experiment:
- Keep the existing P-VFET filter board (2nd order with Fc= 460Hz)
- add one PO89ZB per channel (these are your wellknown modules, 4th order with Fc= 5kHz), between said existing filter board... and the 1500uF per OS I already added
I hope that way to keep on adressing somewhat the 100Hz ripple that comes out from the existing SMPS (but that's not the point here obviously, when considering these filters), while - and that's the new important bit - adressing further the SMPS HF garbage (apparently it is 200kHz and harmonics) with... the addition of your excellent small filters. This combo might be infact better performing regarding HF (at 200kHz and above) than the N-VFET filter.
Whereas it will have at all a sonic impact at the end is TBD, and only that matters!
I might aswell have to adjust the 20V bias again. YThe positive is I am playing with connections and cables to realise quickly reversible experiments where I can add and retire easily the additional PO89ZB filters, from 0 (current configuration) to 2 per channel... should I want to go that far. Admittely at the cost of long cables for these experiments, cables that would if the outcome is positive become short with filters integrated into the VFET amp.
Let's see first if additional filtering like this goes in the right direction...
Many thanks again for your interest and hints
Claude
That is indeed also a potential solution, something I have considered,comething I could still easily build on an experimenter's board...
Apart from the need to probably tame things at start up (many uF) - as NP did with this filter board- the N-VFET is using a RCRC filter (so 2nd order overall with here Fc = 53Hz). It adds indeed many uF that can also help addressing other concerns while at it.
It would require proper modeling to find out how the various Zin and Zout at the filter's interface interact with the filter itself to design a dedicated filter. And the ESRs of the caps - as the ones of the caps I have already added at the OS board PS entry - have also a clear impact.
But that could take me as much time as thinking about it "roughly", as I did in my post 2 months ago, and use the spare time to... just give it a try.
Therefore, at this stage, to evaluate the sonic impact, I will as next experiment:
- Keep the existing P-VFET filter board (2nd order with Fc= 460Hz)
- add one PO89ZB per channel (these are your wellknown modules, 4th order with Fc= 5kHz), between said existing filter board... and the 1500uF per OS I already added
I hope that way to keep on adressing somewhat the 100Hz ripple that comes out from the existing SMPS (but that's not the point here obviously, when considering these filters), while - and that's the new important bit - adressing further the SMPS HF garbage (apparently it is 200kHz and harmonics) with... the addition of your excellent small filters. This combo might be infact better performing regarding HF (at 200kHz and above) than the N-VFET filter.
Whereas it will have at all a sonic impact at the end is TBD, and only that matters!
I might aswell have to adjust the 20V bias again. YThe positive is I am playing with connections and cables to realise quickly reversible experiments where I can add and retire easily the additional PO89ZB filters, from 0 (current configuration) to 2 per channel... should I want to go that far. Admittely at the cost of long cables for these experiments, cables that would if the outcome is positive become short with filters integrated into the VFET amp.
Let's see first if additional filtering like this goes in the right direction...
Many thanks again for your interest and hints
Claude
Last edited:
Triggered by the latest BA vids mentionning the potential use of Papa's last filter for other projects, I had a closer look at the N-VFET filter board. It seems rather that the filtering bit consist of two separate RC modules, and not RCRC. So presumably one RC per channel, each being roughly R=1 and C=3000. So PS channel separation at that stage already, but only first order filter with Fc still being 53Hz...
And there is no soft start device to tame anything at start up, so I presume the SMPS copes with 6000uF (admittely through some R) as expected and that's a positive for my experiment...
Doesn't change my line, but I thought worthwhile correcting my own text above in case someone reads it LOL
Again too quickly typed in the eve...
Claude
And there is no soft start device to tame anything at start up, so I presume the SMPS copes with 6000uF (admittely through some R) as expected and that's a positive for my experiment...
Doesn't change my line, but I thought worthwhile correcting my own text above in case someone reads it LOL
Again too quickly typed in the eve...
Claude
Last edited:
This was a very enlightening and useful post, Claude. Will save a lot of time chasing down some diminishing returns.
Thank you for sharing your findings!
Thank YOU for reading me 🙂
Once I find some time, the possibly last mod (although swaping FEs would be easy in my set up) I will try will be some additional SMPS filters. I shall report as ever.
Have a nice day
Claude
Once I find some time, the possibly last mod (although swaping FEs would be easy in my set up) I will try will be some additional SMPS filters. I shall report as ever.
Have a nice day
Claude
Concerning the output stage, I’ve been doing a fair amount of cap rolling for C4 (0.33uF signal input) and C5 (0.33uF bypass for 10K output cap C1). I have not finished all my trials but I thought I would share some findings.
C5 has larger impact on the sound than C4, but both are noticable and significant
(C1 is a 10,000uF Mundorf in my setup)
For C5 I tried values as low as 0.01uF up to 1K electrolytics.
Lower values like 0.01uF noticably clarified the upper end (I would say sharpened but that might have a negative connotation here) but at the perceved expense of the lower range. If I were bi-amping I would use this value, but otherwise I would not go lower than 0.33uF.
Film and electrolytics from 10uF on up improved the bass slam a bit but not much more than that. The sweet spot for me was between 0.47uF and 1uF.
Cheap polycarbonates (ZM fave) performed well throughout. At 0.01uF they even beat out Jensen silver foils.
My top caps for C5 were Duelund VSF 100V 0.47uF in a virtual tie with Mundorf silver oil 0.68uF with a 1uF polycarb in third. The highest value CuTF I had was 0.22uF, which did beat out the stock 0.33uF. I suspect 0.47 might win the day, but that is for the future.
Different values at C4 did not make as much difference as at C5, but I still preferred between 0.33uF and 0.68uF.
Interestingly, for C4, the Mundorf silver oil 0.68 was top and the Duelund was… terrible, not even as good as the stock. I've yet to try Teflons here so no conclusions yet.
C5 has larger impact on the sound than C4, but both are noticable and significant
(C1 is a 10,000uF Mundorf in my setup)
For C5 I tried values as low as 0.01uF up to 1K electrolytics.
Lower values like 0.01uF noticably clarified the upper end (I would say sharpened but that might have a negative connotation here) but at the perceved expense of the lower range. If I were bi-amping I would use this value, but otherwise I would not go lower than 0.33uF.
Film and electrolytics from 10uF on up improved the bass slam a bit but not much more than that. The sweet spot for me was between 0.47uF and 1uF.
Cheap polycarbonates (ZM fave) performed well throughout. At 0.01uF they even beat out Jensen silver foils.
My top caps for C5 were Duelund VSF 100V 0.47uF in a virtual tie with Mundorf silver oil 0.68uF with a 1uF polycarb in third. The highest value CuTF I had was 0.22uF, which did beat out the stock 0.33uF. I suspect 0.47 might win the day, but that is for the future.
Different values at C4 did not make as much difference as at C5, but I still preferred between 0.33uF and 0.68uF.
Interestingly, for C4, the Mundorf silver oil 0.68 was top and the Duelund was… terrible, not even as good as the stock. I've yet to try Teflons here so no conclusions yet.
My VFET just got a new pair of speakers today. They pair really well with my homage to the LS3/5A in high gloss natural birch finish. Great bedroom amp/speaker setup.
Attachments
Thanks to alazira for sharing listening impressions. Capacitor rolling can be a delicate subject, and it can be rewarding as well as entertaining.
My latest work in that area has been with the Bulwark boards. Each pair uses different signal transformers – one with the original Edcors, and a second with the Jensen JT-123-FLPCH. The electronic gain has been boosted in each set so they match each other closely in overall gain. Both sets now use a 68 pF Wima polypropylene at C7. I have recently switched to using polycarbonate film at the C1 input coupling location. The first set has a pair of 1.0 uF caps in parallel on each board, with the standard 100k values for R6 and R7. The second set has a single 1.0 uF at C1 with 221k Ohms for R6 and R7.
Both sets of boards exhibit an apparent increase in transparency with the polycarbonate coupling caps vs the polypropylene caps listed in the BOM. The blue polypropylene caps are still quite good, and one should not feel at any disadvantage for using those. However, the polycarbonate caps seem to do a better job of "getting out of the way" of the music. Having tried both options with each set of boards, I hear consistent improvement in the form of extra detail of instrumental attack, decay and ambient information from my favorite recordings with the polycarbonate caps in place.
Unfortunately, the polycarbonate film caps seem to be out of production these days, so we are left to search for them online and in our local electronic surplus stores. They are worth trying, nonetheless.
My latest work in that area has been with the Bulwark boards. Each pair uses different signal transformers – one with the original Edcors, and a second with the Jensen JT-123-FLPCH. The electronic gain has been boosted in each set so they match each other closely in overall gain. Both sets now use a 68 pF Wima polypropylene at C7. I have recently switched to using polycarbonate film at the C1 input coupling location. The first set has a pair of 1.0 uF caps in parallel on each board, with the standard 100k values for R6 and R7. The second set has a single 1.0 uF at C1 with 221k Ohms for R6 and R7.
Both sets of boards exhibit an apparent increase in transparency with the polycarbonate coupling caps vs the polypropylene caps listed in the BOM. The blue polypropylene caps are still quite good, and one should not feel at any disadvantage for using those. However, the polycarbonate caps seem to do a better job of "getting out of the way" of the music. Having tried both options with each set of boards, I hear consistent improvement in the form of extra detail of instrumental attack, decay and ambient information from my favorite recordings with the polycarbonate caps in place.
Unfortunately, the polycarbonate film caps seem to be out of production these days, so we are left to search for them online and in our local electronic surplus stores. They are worth trying, nonetheless.
Scourge, Bulwark, Marauder, and Dreadnought are now available as full kits including PCB and all electronic components. Yes even the Edcor transformers.
VFET Front End cards :: diyAudio Store
VFET Front End cards :: diyAudio Store
I think I want to order two definitely the dreadnought which other one do you guys think. Has Mark Johnson or anyone talked about the sonic differences between them.
Well, it is definitely worthwhile to hear one of the designs that includes the Edcor transformer. I can certainly recommend the Bulwark, as my reports have indicated.
Ok i ordered just one of each. I Hope in a month or so there will still be some left in-stock, and then i will order another of my favorite. Is the store limiting the orders per customer.
I wasn't lucky enough to snag one of the VFET kits but any news yet of what other projects these front ends might be suitable for? I have a quad of THF-51S that are itching for a home.
How about this?
Single Ended Tokin SIT THF-51S Common Drain Mu Follower Amplifier , 45W?
There are supposed to be others, but why wait? 😉
Single Ended Tokin SIT THF-51S Common Drain Mu Follower Amplifier , 45W?
There are supposed to be others, but why wait? 😉