• WARNING: Tube/Valve amplifiers use potentially LETHAL HIGH VOLTAGES.
    Building, troubleshooting and testing of these amplifiers should only be
    performed by someone who is thoroughly familiar with
    the safety precautions around high voltages.

6N2P/6CW5 design with toroidal outputs

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Well, since this has been in daily use for the last couple months I figured it was time to declare it a success, and post it up to show off :cool:

Its a variation of my favorite topology, the unbypassed cathode voltage amp, step-network biased concertina, and garter bias. Very similar to my flea amplifier, with a few tweaks. I wanted to go for an all-noval design this time, so went with the wonderful and affordable 6N2P for the front end, and went with rather boring but effective values to set it up. Unbypassed cathode lowers gain and distortion, and the distortion there is will end up primarily 2H from past experimentation with this set of operating points. A little feedback to the cathode is a nice way to tighten everything up ;)

The output stage is where things get a little unusual. For the first time in a few years, I'm actually going with a pentode output stage:eek:. Since garter bias works so well I decided to try it out on here, and it works pretty damn well, even on pentodes that are reasonably matched. I went with the 6CW5/EL86 for this use, with a nod to trying out the 6P43P-E later on (I have 21 of them, may as well ;) )

I went with the Antek AS-0505 power toroid for my output transformers, which works out to 4232 ohms, with both primaries and secondaries respectively in series (for the relevant maths, 115+115=230, 230/10=23, 23x23=529, 529x8 ohms=4232 ohms) which should give pretty good linearity for these affordable and lovely, overlooked tubes. I'm running them a little conservatively at 32mA per cathode, for about 9.2 watts of combined plate/screen dissipation, but will likely increase this soon now that I know this setup works well. Of the tubes on hand, I find the Polam EL86 to be particularly nice.

Interesting to note, that if you connected the secondaries in parallel, and used 4 ohm speakers, these transformers would work well for EL84 or 6V6/6P1P as well, showing a primary impedance of 8464 ohms. Some minor changes to the screen supply would be needed to give a higher range of voltage. Easy, add another zener in the string :)

For power, an Antek AS-1T230 - 100VA toroid gets rectified through a 1N4007 bridge, fed to a pair of IRF820 mosfet filters, one of which uses a string of five 1N5262 50 volt zeners, with a pot/resistor in series feedig the mosfet gate for adjustment range. This gives an effective way of setting the screen voltage. The output tubes get their plate supply straight from the first reservoir cap, and the additional unregulated but smoothed mosfet filter feeds the preamp stage.

All resistors are basic type stuff, concertina resistors matched, and the garter resistors are 10W non-inductive bulk lot wirewounds, selected to within a couple ohms of eachother. All coupling caps are siemens poly films. Most everything I had onhand other than a few power caps, but it's all basic grade stuff for the most part. Used some leftover hammered texture paint for the affordable 12x8x2.5" BUD industries aluminum chassis. I went ahead and sprung for a fancy swooky purple power button LED this time, though :)


How does it sound? Well, very, very nice, if I do say ;)

I'm not one for audiphool jargon, but compared to my flea amplifier it is noticeably different, owing to that pentode sound. Overall it is much more powerful, of course, but it's super clean and accurate, not too clinical or sterile. Sounds like a nice EL84 pentode amp, but with a little more mojo that one seems to find with low impedance tubes. This thing makes thunderous bass, and has a wonderful midrange/high end. Nice general-use amplifier.

Total iron cost, $34.50 power, $17.50x2 for output, plus $13.50 shipping, makes for $83 total. Not bad :cool:


Pardon the terrible picture and sloppy schematics!
 

Attachments

  • IMG_20180518_110247 (1).jpg
    IMG_20180518_110247 (1).jpg
    607 KB · Views: 3,048
  • IMG_20180917_102705.jpg
    IMG_20180917_102705.jpg
    594.1 KB · Views: 3,100
  • IMG_20180917_103430.jpg
    IMG_20180917_103430.jpg
    572.8 KB · Views: 3,062
Last edited:
mondogerator,

The RC network is a voltage divider that drops the DC voltage by half to get the second tridode where it needs to be DC wise to act as a concertina. The capacitor shunts the AC and allows it to bypass the upper resistor on the voltage divider to restore all of the AC gain of the first stage.
 
Sorry, yes. It is to reduce the DC voltage at the second triode grid, while preserving the AC signal level.

In this particular case, the first triode is set up to have ~60-65% of the supply voltage at its' plate, and the divider brings this down to approximately 25% of the supply voltage in order to bias the cathodyne at a good operating point.

Also, a little error, the 1M and 1M5 resistors should be switched on the schematic.
 
I also made a similar amp. It was a guitar amp, but the output stage was tested for power and frequency response before the preamp stage was built. I torture my HiFi amps by playing guitar through them, and likewise I test the amp stage of my guitar amps by playing music through them. This one sounded quite nice in glorious mono and the bass was surprisingly deep and solid considering the use of a similar Antek toroid for the OPT.

The amp ran 45B5 / UL84 output tubes. These are 6CW5's with a 45 volt 100 mA heater. The power supply used a Triad 100 VA isolation transformer. B+ was derived from a voltage doubler and ran about 340 volts. The screen supply was derived from a FW bridge on the same secondary for about 170 volts. The OPT was an AN-509 with both 9 volt secondaries wired in parallel providing a 5200 ohm load. Power at clipping was 22 watts, and the THD was around 3% at 20 watts. A bit more power could be had with a lower load impedance, but I have the Antek already so I tried it. I also tested the amp with a real OPT rated at 3300 ohms. It made 26 watts at clip and under 2% THD at 20 watts......but broke the $100 budget for the amp.

The 6CW5 is pretty good little tube and I have squeezed 30 watts from a pair in my SPP amp boards.
 

Attachments

  • TopView_A.jpg
    TopView_A.jpg
    240.8 KB · Views: 696
  • ToneTest_A.jpg
    ToneTest_A.jpg
    229.4 KB · Views: 2,325
  • Antek_OPT_A.jpg
    Antek_OPT_A.jpg
    250.4 KB · Views: 2,362
Main supply voltage is around ~320-315 volts at the first cap and rectifier, the fixed mosfet ripple filter knocks that to ~300 or so, and this feeds the preamp circuitry. The screen supply is through the separate adjustable mosfet ripple filter/regulator, and is around ~200 or so, depending on tubes and what sort of screen voltage they want, with maybe 75-100 volts of adjustment. The output tube plates take their supply right from the bridge rectifier/cap, before the filters. Any 230V secondary should give you a pretty close raw supply if going for silicon for your rectification, and all the voltages aren't too fussy.

The mosfet filters sure do work well, it's dead silent with no music playing, on my 91db efficient speakers. I'll never buy a choke again with how cheap these are to implement :)

There is a minor error in the schematic, the voltage divider at the concertina grid should be 1M5 up top, 1M on bottom, opposite to what I've drawn. You could also make these both equal at anything from 470k~1M each without hurting anything really.

You could leave out the screen regulator circuitry, and triode connect the outputs too, for a nice 8-10 watts out, but this would of course require some work figuring out new cathode resistor values, or reducing the supply voltage to the outputs slightly. Whatever you need to get dissipation in a safe range, really.

Pretty versatile design, really. One could easily run EL84, 6V6/6P1P, 6P43P-E, or even sweep tubes like the 6AV5GA or similar with changes to cathode resistance, output transformer load impedance, and screen regulation voltage range, as well as triode connect things too. Lots of room to play around. As built with the 6CW5/EL86- pentode connected, figure abut 20-25 watts output.
 
Last edited:
and triode connect the outputs too

The 6CW5 and its higher filament voltage brethren really don't like 300 volts on the screen grid. It may work for a while if run at full tilt all the time, but G2 will glow itself to death when the amp is idle. Max spec is 200 volts, and I keep it lower than that if bending the plate voltage spec a bit. Max plate voltage spec is 250 volts, but 340 works fine if the screen is at 170 volts. I have been to 400 volts without issue with G2 in the 130 volt range. 400 volts is the limit of the small power supply I had at the time of all the 6CW5 experiments.

If you want to triode this amp, stick EL84 / 6BQ5's in it.
 
I'm not a huge fan of triode-connected pentodes unless they're really special, so my tendancy is to go pentode mode and leave things a bit lean for longevity. If going triode mode I think simply using the screen regulator for the plate supply with a big heatsink would be a nice way to go, if a bit hot and wasteful.


That reminds me I need to play with the (twenty-one!) 6P43P-E that I have on hand. When I first tried them in this amplifier (as built) They redplated quickly, so I think that some experimenting needs to be done. They pulled enough current to put over 25 volts across each 500R cathode resistor for 50mA each plate current (280 volts across each tube), and blew up one of the 50 volt bypass capacitors. Redplated at 14 watts?! This was even with the screen voltage set low so I could adjust it up as needed. I have a feeling they will need a much higher value cathode resistance, or a much lower plate voltage to be happy. They seem to be a pretty close equivalent at lower voltages, but more fooling around is needed to see what their happy place is up high. I could just use a CCS on each one, but, well, I don't want to :)

I haven't tried them triode connected, so I can't comment on them there.



Ideally, i would like to draw up a dual mono PCB design for this amplifier setup using EasyEDA, as it could easily be made up on a pair of 100mmX100mm boards, and there sure are enough suitable tubes to make it worthwhile to have on hand. I'll have to play more with the layout I have sketched up to get something symmetrical and easy to use. Should be useful for all those builds all my friends and family keep asking for :)
 
Last edited:
I'm not a huge fan of triode-connected pentodes unless they're really special

Some of them can sound pretty good. The 6V6 sounds pretty good in triode if you want a low $ alternative to the 45. The 5V6 and 12V6 are still pretty cheap. The DHP 307A triode wired makes a good stand in for the 300B. Squeezing 50 to 70 watts from a pair of triode wired KT88's is cool too, and sounds pretty good as well, but as you say, most are just OK.

triode connected 6P43P

I have never seen these, so I can't say what they will eat.

I have maybe 20 GE 6CW5's that I pulled out of a bunch of HP audio oscillators that were headed to the metal scrappers. I also have about 12 French made rebranded UL84's that I bought for the Hundred Buck Amp Challenge, and a few 16CW5's. All behave in a similar manner, you can feed the plate anything up to 400 volts as long as G2 is around 150 volts. The screen, and maybe the plate will glow if you go over 250 volts on the screen. I never tried triode or UL with any of these.
 
I haven't tried them triode connected

I did, they sound pretty good at 200V 35-40mA with mixed bias (100 ohms cathode resistor not bypassed and about 220K g1 to ground through negative fixed bias supply) and seem to be stable. No harshness with a high load: 10K toroidy OPT. A nice low power PP amplifier, even open loop.

Some curves here:

Roehrenkennlinien

6P43P-E triode connected looks quite nice, now take a look at pentode mode :fight:
 
Last edited:
I haven't, and I've been meaning to try it. Next time I order parts I plan to grab a few more types of mosfets to play with. It'll save a whole tube per channel, so worth a shot!

Well, I have some downtime at work, so started to draw up a PCB for this rig. Goal is to have it be semi-universal, one channel per PCB, and be 100mmX100mm. The idea is to use a separate PSU board, or traditional chokes and caps with a vacuum rectifier. Going for a modular, load it out as you like type of setup.

The connections for the plate and screen can be easily jumpered on board, or brought out to a screen supply and UL taps if desired, allowing for triode, pentode, or UL mode connections. Since the original build used garter bias successfully, I've allowed for that on the board as well, with ventilation holes under the resistors. Again, it can easily be jumpered out if standard cathode bias is desired. I may add the choice for cathode CCS loads if anybody has some good ideas for them, I'm just not sure what to choose, as everyone has their favorite CCS.

I'm considering modifying the cathode setup on the front end to allow LED or resistor/cap bias for situations that may want higher gain, with a resistor underneath to still allow for NFB to be used. Could be handy when using medium-mu tubes up front like the 6N1P, and still being able to run some NFB. It's easy enough to add the pads to the existing layout. The cathodyne setup can be biased with the step network as in the original schematic, or self-biased in the more traditional manner, too.

The front end is otherwise the same as the Flea amplifier and the posted schematic at the beginning of this thread, otherwise. should be a neat little board, and especially handy for a garter-bias, triode mode build with the 6P43P-E, 6P14P/EL84 types :)


Keep in mind this is very preliminary, nothing is labelled correctly yet, and I'll be modifying things as I go a bit more. Yes, I know there a few missing components, and no ground/power connections yet :)

Any suggestions, just let me know!

Alright, I'm in a groove, so to speak. Some more work on the PCB. I incorporated the options I listed for the cathodyne and input stage, and started labeling things. I figured having it in the documentation that you populate either the cathode LED or a cathode resistor/capacitor would mean a single LED footprint would work for both the resistor/cap and LED option, or one could simply use a wire link to run the standard cathode resistor/feedback resistor combination.

I decided to keep the same numbering scheme for things as used on the Flea amplifier, just to make it easy for refferring to the existing schematics I have drawn up on hand, and the options will be labelled differently as needed to show function a bit. Spots where wire links are intended to be placed for standard configurations are labeled "LINK" with a resistor outline on the silkscreen to make it easy to figure out at build time. As before, the intent is for you to wire up the heaters with twisted pairs of wire and jump them from socket to socket off board, but a few extra pads will be provided to make DC heaters easier to implement without jumping wire to existing components. Caps are intended to be mounted underneath the board, but will work on top if desired, everything else can go whichever way you want, although the silkscreening is on the top of the board. I'm assuming that most of the warmer resistors will end up on the top of the board so they can breathe a bit.

Yes, there are holes under the sockets for those of you that can't help themselves and feel the need to shove an LED under each tube, but wiring them up is on you :)

About to leave work, so likely the last one for today (I have limited PC access at home) starting to look pretty good and complete...

More progress. I decided to change over to a ground pour on the bottom plane of the board instead of traces this time, and I'm liking it even better so far. Added the ability to run separate supply for either driver and output, and added a place to use a resistor to drop the main supply to feed the driver from the output stage supply as well, for single supply use you simply populate it with a resistor and a 2P terminal block, or a 3P terminal block for separate supplies.

Toying with the idea of putting a mosfet screen supply regulator (simple cap multiplier type) on the board, but I'm afraid of "scope creep" making things more complex than needed. It may be best to keep that on a separate PCB... I'm still not sure how I want to go for that. I do have plenty of room for one. Hmmmm.

Little in the way of updates, but worked up my signature into the board (heh) along with my email address. I'm a huge Tolkien nerd, so my username in elvish is my usual signature for most DIY projects.

Well, I did a little more screwing around with he board. I've incorporated the option to run an LM317 as the cathode load as a simple CCS by request of several of the folks interested. To use it you simply populate the LM317 and it's current set resistor, and jumper out the bottom garter resistor. While not perfect, it is a pretty popular option and should be a cheap and cheerful way to go for those of you who would like to use it :)

I'm thinking that this board is pretty close to being done as far as the general idea goes. I'm not sure what (if any) other features would be useful, or even fit :cool:
 

Attachments

  • PCB_NOVAL-MONO_20181114102940.png
    PCB_NOVAL-MONO_20181114102940.png
    84.4 KB · Views: 242
  • t0RF3LVfWFX.png
    t0RF3LVfWFX.png
    80.1 KB · Views: 268
  • PCB_NOVAL-MONO_20181114135211.png
    PCB_NOVAL-MONO_20181114135211.png
    156.7 KB · Views: 250
  • dLWtbSx16St.png
    dLWtbSx16St.png
    84.1 KB · Views: 252
  • PCB_NOVAL-MONO_20181114153649.png
    PCB_NOVAL-MONO_20181114153649.png
    172.7 KB · Views: 2,724
  • wq1MJwKYse9.png
    wq1MJwKYse9.png
    85.1 KB · Views: 2,911
  • PCB_NOVAL-MONO_20181115100655.png
    PCB_NOVAL-MONO_20181115100655.png
    225.4 KB · Views: 623
  • EpppBNWp5bW.png
    EpppBNWp5bW.png
    85.3 KB · Views: 583
  • 9faatxz4br2.png
    9faatxz4br2.png
    70.8 KB · Views: 563
  • PCB_NOVAL-MONO_20181120135156.png
    PCB_NOVAL-MONO_20181120135156.png
    331.1 KB · Views: 213
  • Screenshot_2018-11-20-13-53-41.png
    Screenshot_2018-11-20-13-53-41.png
    167.7 KB · Views: 197
  • top122718.png
    top122718.png
    90.8 KB · Views: 193
  • bottom122718.png
    bottom122718.png
    81.2 KB · Views: 195
  • PCB_NOVAL-MONO_20181227110542.png
    PCB_NOVAL-MONO_20181227110542.png
    242.6 KB · Views: 221
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.