another 555 smps layout

[Marcos - Munky]

A few days ago, there was a bit of talk about the 555 "nixie" smps on the pictures topic. So I like to share what I've been using to power my tube preamps.

Basically, I'm using that schematic that's been online for some time. The circuit works very nice, but from what I've read the layout can be a bit tricky, and indeed I noticed a some differences while I was testing a few different configurations. One of them, with random parts placement, was very noisy. Then I've read the IC, inductor, diode and mosfet should be as close as possible to each other, and started to modify my layout.

Another thing that I kept in mind on my tests was the mosfet temperature and placement. I saw a good number of layouts with the mosfet at about the middle of the board, with no place for a heatsink. Some people said they set the smps for a higher voltage and/or powered a 2 tube preamp and the mosfet was cold or just warm to the touch, but on some of my tests I got a kinda hot mosfet. I tried a few different models. One of the main specs for the mosfet is the Rds, a lower Rds means a cooler mosfet. But most of my mosfets came from aliexpress and they do work on this circuit, but they may be remarked/out of spec and lots of places are selling the same kind of mosfet (and a lot of other suspicious parts), so we can't be sure if we're getting the real deal. For that reason I made my layout with the mosfet closer to one edge, so it can be mounted on a heatsink/enclosure if needed, which also works to hold the board in it's place. Just keep in mind the mosfet must be insulated from ground, or else it'll burn.

Here is the latest layout I've made. I added holes for different kinds of trimpots. And I added a extra pad for voltage input, so I can just run a wire from the power jack to the smps board and a wire from the smps board to any other circuit I'm using instead of connecting everything to the power jack. I had to add a jumper, and this jumper runs under the 555.

While I didn't took any current measurements, it powered with success a single tube preamp (Alembic F2B), a single tube preamp using the EF184, the JCM800 preamp (2 tubes), the Mesa Boogie preamp (3 tubes) and the Superfly (submini tube amp). It can't power stuff like tube amps using regular size tubes.

The circuit and layout are pretty quiet for almost everything I powered with it. I tested it with a 12V bench power supply, a 9V pedal power supply, and even a 5V PSP charger. On the EF184 build, I started with a 19V laptop charger to power a class D amp, dropped it down to 6V using a DC-DC converter to power the tube heaters and used the same 6V to power this smps. Even with all those power convertions, the result was a noise-free amp.

And finally, here's a photo of the finished smps board:


Hope this layout helps anybody who like to give a try on a high voltage preamp.

[temol]

Nice and compact layout. I would leave a bit more space between inductor and transistor. Diameter of the inductors varies (depending ong current rating). Also, U shaped heatsink may collide with the inductor. Do not forget about mounting holes on a pcb

I run mine @ 290 to 320V, depending on the preamp. What's your usual output voltage?

[marcelomd]

Nice!

Do you know the current output?

Thanks!

[Marcos - Munky]

Temol, I mount the mosfet on the enclosure even when it runs cold, so I don't really need a mounting hole . I surely can add some holes and also mod it for a bit more space between inductor and mosfet, but for some reason my eagle isn't showing any holes anymore. I messed up with something, so I'll have to try to reinstall it and see if it works again. For voltages, I usually set them by ear, but it's pretty rare to for me to go that high. I'd say I stay from 220V to 250V, which works fine for dirty preamps.

Marcelo, the current output depends on a few things, like the set voltage and the inductor. I didn't measured the current, but it's safe to say it's enough to power a 3 tube preamp/effect. Maybe a 4 tube preamp. But you can't power a tube amp with it. Well, you can power a amp with a single 12AX7 on the preamp and a 12AU7 as the output power, like the Firefly, but this is probably the limit of this circuit.

[marcelomd]

Marcelo, the current output depends on a few things, like the set voltage and the inductor. I didn't measured the current, but it's safe to say it's enough to power a 3 tube preamp/effect. Maybe a 4 tube preamp. But you can't power a tube amp with it. Well, you can power a amp with a single 12AX7 on the preamp and a 12AU7 as the output power, like the Firefly, but this is probably the limit of this circuit.

I was thinking about preamp tubes, really. I'm done messing with big transformers. By 3-4 tubes you mean 6-8 triodes, right? More than enough =)

[Marcos - Munky]

I was thinking about preamp tubes, really. I'm done messing with big transformers. By 3-4 tubes you mean 6-8 triodes, right? More than enough =)

Yes, 6-8 triodes. I tested it with 2 6N2P on a JCM preamp and 3 6N2P in a Mesa Boogie preamp. The 6N2P needs a bit more current than a 12AX7, so it'll probably handle 4 12AX7 when used as preamp stages. With "preamp stages", I mean somebody from this forum (I think it was Nathan) tried to power a tube compressor using 3 12AX7 and had some issues, but the tubes was used in a configuration that requires more current than on a regular preamp stage.

And 6 triodes is exactly what is needed to build a Fender-like clean preamp plus a JCM preamp.

[vigilante397]

I was thinking about preamp tubes, really. I'm done messing with big transformers. By 3-4 tubes you mean 6-8 triodes, right? More than enough =)

Yes, 6-8 triodes. I tested it with 2 6N2P on a JCM preamp and 3 6N2P in a Mesa Boogie preamp. The 6N2P needs a bit more current than a 12AX7, so it'll probably handle 4 12AX7 when used as preamp stages. With "preamp stages", I mean somebody from this forum (I think it was Nathan) tried to power a tube compressor using 3 12AX7 and had some issues, but the tubes was used in a configuration that requires more current than on a regular preamp stage.

And 6 triodes is exactly what is needed to build a Fender-like clean preamp plus a JCM preamp.

The tube compressor was 2 12AU7 and a 12AT7, both of which pull a lot more than 12AX7 on the plates. Still not convinced that was the problem, I think I was doing something else wrong, but whatever, I redesigned it down to two tubes anyway I've used this SMPS for 4 6N21B, which pull about the same as a 12AX7, so I can verify it should be able to do 4 12AX7.

[Elijah-Baley]

Great work, Marcos - Munky.

I have a problem: I can't etch board.
After what you said about the parts placement I'm wondering if I have to realize, copying your layout, a protoboard of it successfully. In this case, I think I need some method advice, because, though I realized two protobards, I failed the first tests, before make them work fine. Failed this can be more dangerous.
Or I have to draw a veroboard layout, trying to copy your layout. Stripboards are easier for me, but the layout will be unverified until I tried it and nobody can gurantee it will be good as the Marcos - Munky's original PCB.

What do you think about it?

[rankot]

I will post my layouts and measurements in following days. Thanks for starting this!

[Marcos - Munky]

I've used this SMPS for 4 6N21B, which pull about the same as a 12AX7, so I can verify it should be able to do 4 12AX7.

Thanks for the confirmation, Nathan!

After what you said about the parts placement I'm wondering if I have to realize, copying your layout, a protoboard of it successfully...Or I have to draw a veroboard layout, trying to copy your layout. Stripboards are easier for me, but the layout will be unverified until I tried it...

The things with parts placement are the smps can be noisy and have less efficiency. If you're doing it just for a quick test and not pushing it near the max current output, that's not a problem. It'll still handle the current needed, and a protoboard build or a test outside an enclosure can be noisy anyway . Copying the layout from pcb to veroboard should be easy, but if stripboards are easier for you, there's this layout, based on the same circuit.

I will post my layouts and measurements in following days. Thanks for starting this!

Nice! I'll also try to do some measurements in following days. And will update the layout with mounting holes and a bit more space between the mosfet and the inductor as suggested by Temol.

[Elijah-Baley]

Thanks for the layout linked.

I can't find the UF4004 in none of the shop where I used to buy. I read the 1N400X is not the same thing. What can I use?

[rankot]

Thanks for the layout linked.

I can't find the UF4004 in none of the shop where I used to buy. I read the 1N400X is not the same thing. What can I use?

Any similar ultra fast diode - UF4004/4005/4006/4007 will handle up to 1A average current, but 4004 has better reverse recovery time than it's "older sisters" - 50 vs 75ns. So if you want something similar, aim to those parameters, and keep in mind that your diode must handle proper reverse breakdown voltage of at least B+ needed (with some headroom), too.

You may try:
DO27> SF36/38 (35ns, 3A, 400/600V), HER307 (70ns, 3A, 400V), HER304
DO41> EGP10G, GP10F, MUR260, MUR140, STTH2R06, STTH108, SBYV26C, BYV26, BYV27/600...

I also wanted to try IRFP460 (Rdson=0.27Ohm), but it's much bigger than usual IRF740.

[vigilante397]

Thanks for the layout linked.

I can't find the UF4004 in none of the shop where I used to buy. I read the 1N400X is not the same thing. What can I use?

I've used 1N4936 from Tayda, they work just fine for these as well.

[Elijah-Baley]

Thank you.

Tayda has the UF4007, and the reverse recovery time is 75ns.
https://www.taydaelectronics.com/datasheets/files/A-3297.PDF

In the 1N4936 the reverse recovery time seems very high, instead. (Over 150ns)
https://www.taydaelectronics.com/datasheets/A-487.pdf

Like for 1N4937.
https://www.taydaelectronics.com/datasheets/A-488.pdf

I found BAV21 with a reverse recovery time of 50ns, but the peak reverse voltage is, maybe, a bit low: 250v.

And the MUR120RLG, with a reverse recovery time of 35nF, but Peak Reverse Voltage of 200v is too low.

I don't know what's difference make this reverse recovery time, but I think the better choice on Tayda is the UF4007.

[vigilante397]

When I tried the 1N4936 Tayda wasn't stocking UF4007, good to know that they do now. A typical NE555 SMPS only switches around 50kHz or so, which puts it out of audible range, but it still doesn't need an insanely fast rectifier. Converting frequency to period gives us 16,000ns at 50kHz, so if the diode's recovery is significantly lower than that then it should work. As I said the 1N4936 worked just fine, there were no performance issues with a slightly slower rectifier.

[Marcos - Munky]

I can't find the UF4004 in none of the shop where I used to buy. I read the 1N400X is not the same thing. What can I use?

I used UF4007 in my builds, it's the only UF diode I could find at the local stores and I didn't even searched for another alternative.

I also wanted to try IRFP460 (Rdson=0.27Ohm), but it's much bigger than usual IRF740.

If you're ok with lower-ish voltages, you can give it a try with the IRF640 (Rds=0.18Ohm and Vds=200V) or the IRF644 (Rds=0.28Ohm and Vds=250V). I've used both before. You just need to be aware they can handle less voltage than the IRF740 (Rds=0.55Ohm and Vds=400V) and the IRF744 (Rds=0.63Ohm and Vds=450V).

In my experience, 180V to 220V are already enough for a tube amp to sounds good. A higher voltage may be very welcome and is very achievable with the IRF740, but it may need a heatsink. Lowering the voltage on the IRF740 or using the IRF640/644 may give the same result, a cooler mosfet and a lower voltage.

I didn't solved my eagle not showing holes issue yet, but I'll install eagle on a old laptop I have here, so I'll probably update my layout with Temol's suggestions later today, along with some voltage measurements (no load, for now).

[Marcos - Munky]

I updated the first post with the link for the layout with Temol's suggestions. Also, one I have already built gave me a voltage range of 98V to 275V. It didn't gave me a voltage as high as Temol's, it's probably because of the inductor and/or mosfet I've used (I'd bet my mosfet is fake ). But it's still more than enough to power a tube preamp.

[Elijah-Baley]

I'd like to ask something about the amount of the voltage in a tube preamp.

I don't want to go off topic, I'm not sure if it's so simple. Can I power up a tube preamp how much I want? 150v, 200v, 250v... What's change? Headroom, Gain...

[Marcos - Munky]

It changes headroom. Dirty circuits like the JCM800 preamp may sounds better to your ears on low-ish voltages. Eddie Van Halen used a variac to lower his amp's voltages - but he said on an interview he actually rised the voltages, which ended up with lots of people blowing tubes trying to do this . Even on transformer powered preamps it's kinda common to see some of them operating on a low-ish voltage like 170V when you want them to distort a bit.

You have to take a look on the datasheets and check the maximum voltage a tube can handle on the plates. A 12AX7 can take 300V on the plates and you can see a range of 100V to 250V under typical operation, but it can go lower than this value. A 6N16B have typical operation of 100V and max voltage of 200V.

It's kinda like solid state circuits. You can rise the voltage a bit, but not too much or else something may release a magic smoke. You can lower the voltage, but not too much or else you'll get that low voltage misbiased circuit sounds.

[vigilante397]

I built a Marshall 18W amp from a really cheap kit several years ago, and the supplier put the wrong power transformer in, so instead of getting 270VAC I was getting about 500VAC from the secondary. The amp sounded ABSOLUTELY AMAZING up until the tubes melted. I never measured what B+ I was getting, but I'm sure it was absurdly high.

So generally I try to run preamps as high as I can get them without getting ridiculous. 6N21B I always ran at 200V, 12AX7 I currently run 235-240V because I use 250V caps on my builds. But as Marcos said it's a preference just like anything else, lower voltage will change the way it breaks up.

On one design (which I've since changed) I actually took the typical 5k trimpot from the SMPS and used a 5k pot mounted externally on the box so you could adjust the SMPS voltage while playing. It worked fine, I added a resistor on the end so the sweep went from 30V to about 180V, but unfortunately the SMPS switching frequency gets too low at lower voltages so it's in the audio range and causes noise on the line. In asking people that tried it what voltage they liked, most agreed that they liked it all the way up (180V), but there were a couple who said they preferred it around 75V.