ZVS boost converter tube amp

[Zozobra]

Zero Volt Switching SMPS boost converters have been a big step for folks who want to make their own seriously high voltage toys such as Tesla coils and induction heaters for smelting. They take in a fairly pedestrian DC voltage and spit out 100's of Volts to up to kilovolts depending on the device which can then power a flyback transformer to get super high voltages (10's of kVs ). On ebay there is a nice little 75W unit that will take 8-36V and deliver up to ±380VDC and it only costs $6. This is super handy if you have a bunch of old laptop power supplies and they will deliver more power than we need and are well regulated and thus can deliver a nice solid HV supply. There are DC DC buck converters that work on a similar principle that can be used to lower the supply voltage for super clean DC heater supplies too and they are even cheaper.

So I have picked up both a ZVS boost and a buck converter to build a full featured mini amp for home use. These things are capable of much more power than I'll be using here and with a suitable DC power supply then a 30-40W PP amp could be built and the power supply would only cost 7 or 8 dollars, assuming you have an old laptop PSU doing nothing!

Here is the ZVS running off a shitty unregulated 9V@500mA wallwart to test the unit:


Boom, works as expected. There is a little trimmer between the DC input and filter cap on the left of the board which allows you to dial in the output voltage from anywhere between just above the input to the full 380VDC. I must add a quick note on safety here and state that if anyone does want to go down this route that it is absolutely imperative that the PSU that they use has the safety ground pin on the plug connected to the DC negative terminal so that the chassis can be earth bonded to stop the chassis from going live in the event of component failure of an HV short to the chassis! This should be common on all good laptop supplies but you must check first. Shitty 3rd party ones often use a floating supply and aren't suitable for use here as there is nowhere for the HV to go if something goes wrong.

This thread and my Imgur album will be documenting what can be done with these cool little supplies. I've picked up a 600x400mm sheet of aluminium as a test chassis to try out circuit elements in without having to worry about space or having to care that I've drilled a million holes in it. Some folks will probably remember that I built a small ecc99 amp in a purple Hammond chassis while back, well originally my plan was to prototype using a DIY aluminium sheet chassis but I fucked up bending the sheet as my DIY brake left a lot to be desired. I then used the Hammond chassis as I wanted to get on with it, although it was nowhere near big enough. This time I'll be using the metalworkshop at work to do the bending so it should be smooth sailing.

In the end I want to shoot for something like the Fryette GP/DI in terms of functionality in so much that as well as being a standalone 4W amp I want it to have an FX loop (MOSFET a la metro zero loss) and an internal reactive dummy load for direct recording. Since I'll have a solid low voltage DC supply to use that wont fuck with the HV supply I can test out adding a buffered isolation transformer to the output of the dummy load and perhaps even an analogue cab sim for hooking straight into the desk. The actual preamp doesn't matter so much at this stage but I want something blackface fender and a modern high gain channel as a starting point.

This will be by far the most ambitious project I've ever undertaken so don't expect particularly fast progress I'll have to get my hands dirty etching some PCBs to do some of the things I want to do when it comes to time to do things properly but that is a way off. Time to learn how to use DipTrace or KiCad I guess....

[Ry Manchu]

Really cool build idea. It would be good for someone who lives in a boat.

I would do H&K Triamp, Cornford RK100, Bogner XTC, Splawn Nitro, Fortin Cali or Engl Savage.

[Zozobra]

I guess boats have DC power supplies? 28VDC like in aviation?

My first port of call is something that looks a bit like a VHT/Fryette design which is to say that the last stage has local NFB so it cleans up. I do like the Nitro too so I'll be trying that out no doubt.

If this works out then the ZVS boost supply is a pretty simple way to make preamp pedals and could be a good way for folks to get their feet wet without dropping a ton of cash.

[ajaxlepinski]

It's amazing that such a small device can step up 9v to over 300v!

Good luck with the project. Looking forward to pics in progress!

[Ry Manchu]

Most new boats are 24v, most old boats are 12v. They use deep cycle batteries for self power.

[Zozobra]

Yeah it's pretty cool. I was looking at building an SMPS boost converter that a lot of folks use in preamp pedals but these are better in pretty much every way it seems. It opens the door for some cool stuff to be built and having absolute control over the voltage at the first filter node is seriously awesome. I have a few small power transformers that will be utterly redundant if these work as expected as if nothing else, I can work at the voltage I want, rather then having to fudge it.

This dude built an EL34 SE amp using the same board: https://drive.google.com/open?id=0B2sbX ... jhGVlpLZFE


I'll also be abusing a lot of MOSFETs here; a full MOSFET FX loop and possibly an LND150 LTPI just for the fuck of it. I may also add an LND150 boost module right at the front so either channel can get a nice clean boost, although it'd have to be better than any of my boost pedals to make it to the final revision.

[Racing]

SMPS might very well become an option to me when we can design a few more parameters in an easy operable manner.

At this point tho i for one think you should run the LND as PI. Nothing new per se,the LND apart maybe. For highgain use Adam Laboga went down that route with his RAD3. Check it out. (Adam knows his shit)

[Zozobra]

Awesome, I'll see what I can dig up on it. Was it a Cathodyne or LTPI type do you know? I've got a few solid designs on using enhancement mode MOSFETs as a Cathodyne. They're pretty much the same as the tube version, just with a divider from the B+ to gate/ground to bias. I'd like to try an LTPI version so that I can add the NFB directly to it. If I use a Cathodyne type I need to apply it to the preceeding stage, which would likely be an LND150 FX loop return stage, and I want to apply local NFB to that stage to get the input headroom up to 3 or 4 Vpp so that it won't clip, but will still have a healthy output. I don't need much drive to push ecc99s after all!

[Zozobra]

The aluminium sheet turned up this morning. I plan to drill out/pilot the front and rear panel holes tonight so that I can bring it in to work tomorrow to fold into a chassis. I'm gonna fold it with 60mm front at rear panels so I have a huge 600x280mm inside to play in. Much easier to work in than the hammond chassis I was using!

[Zozobra]

Managed to sneak into our workshop today to bend the chassis.
Approx 600x280x70mm so there is an insane amount of space to play in.

[Racing]

Yep.
That a big MF. Right there.

[Zozobra]

I want to try out all sorts of dumb shit so I gotta have space to play

[Zozobra]

I'm starting with a Laney Lo-input-esque 3 stage + CF topology. I've been wanting to try out local NFB ever since I worked out how it works in Fryette designs so I've wrapped a local NFB loop around the last stage and CF. It's variable that when the pot is at 0, maximum NFB is applied and the stage will run clean so long as the overall gain isn't so high it collapses the loop (green trace). Turning the pot up will reduce the clipping threshold until eventually the stage clips as if the NFB wasn't there. Traces and pot rotations are (traces are taken from the output of the tonestack):
green - 0%
blue -25%
red - 75%
cyan - 100%

This should allow the clipping threshold to be nicely dialled in to taste so that either the amp runs nice and clean or the amp runs dirty and cleans up on the volume pot nicely. It also should mean that the amp can be set just on the clipping threshold and then switching on a boost/dirt pedal will dirty it right up as a small push in signal swing will tip it over. The NFB sort of acts like a compressor of sorts too so there wont be a huge jump in volume between the clean and dirty tones. I have no idea if this will work as planned

[Racing]

I´ve actually worked with the ECC-99.
In my case as a PI in an old Schaller KV-50. The earlier all tube version. As you probably already know the 99 can push a fair amount of current,seing what it is.
However.... Triodes DO present a different sound palette than does pentodes why i at least would do that there that you have in mind with one of the old small signal pentodes instead. Like the 6SJ7 or similar. But hey...that´s just me.

Then.. Don´t be afraid of fooling around with "different" tubes as far as preamp. Look at how you set the entry stage up. Amplification factor will become limited,of which i agree,but as such maybe you should set it up in such a manner that you can easily swap for whatever..
Take my word for this,tubes are tubes. Sure,triodes vs pentodes vs tetrodes aso differ. All good,however..for various reasons the general masses seem to believe that the 12AX7 is the only way to go seing its inherent amp factor (mu) of 100. In reality we rarely set them up to produce more than 50-55 times and seing that... It can in many instances be a wise move to have a look around instead.

To NOT maximize amp factor out of the entry stage is IMO a rather smart move. Ditto,please try different input impedances. This will have a rather profound impact on hiss and hum...
For instance. When i use small signal pentodes (that carries with them an inherent amp factor that is WAY WAY WAY bigger than the one of triodes) i rarely use gridleaks of more than 220k. In fact i´ve come to use all the way down to 47k...and it works well. Granted that triodes have the Miller effect and what not to take into account but..none the less...

Just food for thought.

[Zozobra]

I have a few other novals to play with too. I've done some toying with an EF86 and I have some AT and AUs on order to try out some designs with
From trying out a few circuits with the ecc99 I really quite like it. It sounds really big for a small tube and the tone isn't limited by the lack of screen compression effects and the dual triode package is super convenient. I want to try building a mini model t too (which is to miss the point somewhat I know) and here I think a triode power amp will be closer to the ultralinear operation they used then a pentode PA would be.

I need to strip the OT and other bits out of the old chassis to be repurposed here. I probably wont get them installed until later in the week as I'm going to see crowbar tonight and need to get my shit together shortly.

I've got the DC input and the converter board installed in the test chassis today:




Ah the slow grind of metal work...

[Zozobra]

Wired up the LTPI and power amp board today. Metal work and fitting/wiring tomorrow.

[Racing]

Impressive. Have to say,credit where credit´s due.

On the whole SMPS idea i can´t but agree from many a standpoint. In reality,to me,the issue more lies with the LACK of a powertransformer. Powertransformer sag is,at least IMO,a REAL big deal as far as the total package when talking tubeamps. When we´re talking conventional ("normal") tube amps this can amount to 100VDC at the extremes...
What i´m saying? Well,when we endeavour paths like this we need to keep the reference in mind. Power transformer sag is certainly part of that picture as i see it. Could it be catered for in a SMPS unit? Certainly. Absolutely positive about that,it´s just that then...it´d STILL be uniform.

If we strive aside a bit and let for instance Nutubes enter the same recepie the next Q arise and that is the actual NEED of a SMPS.? Or? The Korg versions,thus far,wants to see approx 5VDC. Granted that they thus far don´t really cater to "us" but they will... (Vox just recently presented their "toy" in all of this for instance)
In turn,dunno how into it you are,but recently Berkley announced diodes made from aromatics,and these can be "tailored" behavior wise. Ie;when used for diode clipping for instance their sonic behavior,for lack of better words,can be tailored to suit a certain goal. Ie;it can according to them mimic "tube behavior". There´s videos on YT on the matter.

In short..there´s shit going on since a while back and at least i for one believe we need to keep a keen one eye open. Two ears and one mouth as the saying goes.

Back to your SMPS tube amp then.
No. Triodes will never mimic a pentode,even when UL. Be aware that the UL principle is a AC ordeal FIRST. It is NOT about the DC voltage inherent.
In turn all triodes,at least to my knowledge,carries with them the Miller capacitance no matter what we do. Something to at least be taken into account IMO.

Back to SMPS tho.
I use SMPS for running optos,any and all reverb "bricks" and what not these days. It makes the setups interference free and completely without drama. IMO those that actually BUILD tube amps and are not using SMPS to their advantage are stuck in the 20th century. Yes,it works THAT well.
In turn something as simple as SMPS run heaters of a tube amp would,all other being equal,make for cheaper,lighter and more to the point PT´s. Win-win IMO seing how small SMPS can be had and STILL hold the oumph needed.

[Zozobra]

Thanks. I'm not so concerned about the lack of sag, for this application at least. As you say, sag isn't a big factor in triode power amp operation since sagging the plates won't really change the operation any great deal. If I were using small pentodes in push pull I'd add in a sag resistor before the main filter cap to simulate a bit of PT resistance since the screen effect would be worth it most likely. Will it behave like a linear PSU? Probably not, but it doesn't matter so much for the intended application. This said, in some of the more exotic custom high gainers some folks run 500-520VDC on the plates and then regulate the screen node (and subsequently the preamp) down to 380-400VDC to improve tube life and maintain punch at low tunings. It probably wouldn't sound so great for a JMP though really, this is just a bit of a goof off and the SMPS board offers a lot of off the shelf flexibility to try out different supply voltages and I like to try new things.

The main thing I want to try with this is building a few hybrid circuits. For ultra clean and super high gain moving to a solid state poweramp could be beneficial and a 100W class D mono board costs all of $10. It could also be cool for converting old shitty SS combos into something more tonally pleasing. The other cool thing is that it would be easy and cheap to build small tube preamps as a well regulated 12VDC laptop PSU will do the job just fine, thus the major expense then becomes the chassis/enclosure and there is no need to dick about with cutting IEC holes. Lots of folks over on diystompboxes have been homebrewing their own SMPSs for this application and they work just fine. I don't see the benefit in etching my own boards and sourcing appropriate inductors given the low cost and availability of these boards!

[Racing]

One hundred percent agreed on the preamp thoughts of yours.
As them triodes normally work in true class A at all times no matter what really...yeah well..

[Zozobra]

For the most part, that'll be what I end up using them for. It's be great to have a few different pres kicking about for recording with and the cost will be minimal.

What I really need is a weekend without 200 social engagements so I can finish this test rig up

[Racing]

For contrast and SnG.

[video]https://www.youtube.com/watch?v=AF8kA1t6FTk[/video]

That there is the exact opposite. It´s a -34 vintage 3 watt Philips radio,that for the most part still holds its stock tubes.
Let´s just agree on that they were handed a task..that´s different than the original idea.

Yup. Old 4V heater unit going metal. With a POS chinese Squier and an MXR OD.

Funny tidbit is that the stock onboard AlNiCo speaker actually still works!

[ajaxlepinski]

Thanks. I'm not so concerned about the lack of sag, for this application at least. As you say, sag isn't a big factor in triode power amp operation since sagging the plates won't really change the operation any great deal. If I were using small pentodes in push pull I'd add in a sag resistor before the main filter cap to simulate a bit of PT resistance since the screen effect would be worth it most likely. Will it behave like a linear PSU? Probably not, but it doesn't matter so much for the intended application. This said, in some of the more exotic custom high gainers some folks run 500-520VDC on the plates and then regulate the screen node (and subsequently the preamp) down to 380-400VDC to improve tube life and maintain punch at low tunings. It probably wouldn't sound so great for a JMP though really, this is just a bit of a goof off and the SMPS board offers a lot of off the shelf flexibility to try out different supply voltages and I like to try new things.

The main thing I want to try with this is building a few hybrid circuits. For ultra clean and super high gain moving to a solid state poweramp could be beneficial and a 100W class D mono board costs all of $10. It could also be cool for converting old shitty SS combos into something more tonally pleasing. The other cool thing is that it would be easy and cheap to build small tube preamps as a well regulated 12VDC laptop PSU will do the job just fine, thus the major expense then becomes the chassis/enclosure and there is no need to dick about with cutting IEC holes. Lots of folks over on diystompboxes have been homebrewing their own SMPSs for this application and they work just fine. I don't see the benefit in etching my own boards and sourcing appropriate inductors given the low cost and availability of these boards!

It could also be cool for converting old shitty SS combos into something more tonally pleasing.
Is this something that would work in a Randall RG80? I love my RG80 but, there is always room for improvement.

[Racing]

Yes,but it´d take a bit of understanding on the subject. Designing circuits that is.

Also be aware what i mention above with diodes from aromatics. My bet is that this´ll hit hard the day they become commercially available as most,if not all,distortion produced by SS amps is from diode clipping.

[Zozobra]

It could also be cool for converting old shitty SS combos into something more tonally pleasing.
Is this something that would work in a Randall RG80? I love my RG80 but, there is always room for improvement.

I'd leave an RG80 well alone. Great amp! Here I'm talking about finding some shitty pawnshop/craigslist SS combos and then stripping it down to just the power transformer and a bolt on rectifier+cap to then use as either a total low power tube conversion (if the PT is big enough), or to run a tube front pre and a then throw in a small SS power amp board. I plan on doing this at some point as a grab and go combo to take to jams and I'll no doubt write it up here.