SET amp from 1619 driving a 809. Opinion wanted.

I have had a need for a flea power, very high quality stereo power amp now for the purpose of evaluating my high efficiency full range loudspeaker designs for suitability of working with customer's 45, 2A3 and single 300B type SET amps in the marketplace. I wish to be DC coupled and employ only DH tubes for their alleged magic tone. DC heat will be employed throughout.

Since my inventory of tubes is not blessed with such valuable gems as 45's or 2A3's, 6B4's, or DH input suitable tubes of conventional choice I have decided to be creative from what I do have in stock.

Following the general theme of the famous WE 91A amp that used a pentode to drive a 300B, I have decided to use a 1619 as the line input stage and driver. This transmitting beam pentode tube is similar to a metal 6L6 with a 2.5 volt thoriated tungsten DH cathode. I have 11 NIB, NOS. I was torn to use one also as the output tube in triode mode and still may. This would be about the equivalent of a 2A3. I am however looking first to use an 809 power triode because of the esthetic appeal. With so much effort being put into making this amp look like art (not being able to just throw stuff together is a curse of mine), I would like to see the friendly, soothing, reassuring glow of DH valves. The reward for all this effort!

My circuit directly couples the driver plate to the output tube grid with a plate choke on the driver (reactance drive). My concern is what will happen when the 809 gets into A2 operation. I don't know how well it will perform kept in A1 land only. Unfortunately I have no interstage xfmer suitable to drive the 809 into A2 properly and I cannot wire the input tube as a follower as I need the stage gain and do not wish to go to 3 stages.

Attached is a picture of the layup so far. The top plate is polished 3/32" yellow brass under the protective paper. Wood is oiled black walnut.

Has anyone employed 809's in SET, especially A1? What were your results please?

Yes, I am worried that it may not work in A1 that well. I am using one or other of a 15 watt Hammond vintage stock P-P transformers. I have a set of NOS each of 1610 (10K CT and separate 4 ohm and 16 ohm speaker windings) That gives me options of half or full primary and using the 4 ohm winding in cathode NFB of the 809 if I wish. 16 ohm to 8 ohm speaker load which would give me 5 K or 2.5 K on the anode.

The other two xfmers I have are 1615. Those are 5 K CT and have 4-8-16 ohm taps. I can get a large variety of AC plate load options with these from 2.5K to 20K.

I am parafeeding the ouptut xfmers with a series motor run cap so no DC in the OPT core. A plate choke is used. The coupling cap is also to be connected in ultrapath. I have an idea of running about 40-50 mA on the 809's at about 470-500 volts plate to cathode (~20-25 watts PD.) The tube is rated at 25 watts CCS.

The 1619 driver sits underneath the cathode of the 809 and will see about 250 volts plate supply through the 809. Total B+ should be up around 700-750 volts. Choke input filter for good regulation.

Planning for possible failure, if the 809's don't work out for me I can plop in 811A's and dink around with those, or 300B's with only minor changes to the filament supply and bias resistors. Real SE output iron can also be acquired eventually if the amp looks to be worth spending money on and performance is not yet stellar.

This project is all from the junk box for now except for some big value, low voltage electrolytics I had to order for the DC filament supplies. Shielded filament xfmers are custom by your's truly.

The case was made with the largest pieces of walnut I could salvage. Initially I wanted an 11"x17" chassis. This one is just 9"x10". The circuit is dictated by the amount of wood I had and available parts. No room for two more tube sockets I'm afraid. A pair of IT's could be fit underneath if I make room by getting rid of plate chokes for the ouptut stage. To do this I'll need to buy SE output xfmers for topside. That will cost money. So will IT's. Not an option just now.

This whole thing is just an experiment. If tweaking doesn't work, then it certainly can be made to work when I can throw some parts money at it.

I guess I'm into this far enough now that I'll know the answers to my questions soon enough finding out by myself.

Took me all day but I now have the brass top plate machined and overcoated with clear lacquer. A second coat of linseed oil is drying on the base and tomorrow I can start assembly.

I changed formation from the flying Vee (the tube location pattern first considered) and moved the electrolytic caps in between the rectifier and the output bottles. This should reduce the possibility of induced rectifier noise by more distance and grounded obstacles, and gives me more room to accomodate other output xfmers later on just in case.

I will need to make sheet metal sub-chassis for underneath. I want to be able to lift the entire circuit out of the case by the top plate and not have guts screwed to the insides of the wood. There is just no room to work in there that way. This thing is going to be well packed.

tubelab.com said:

I have some 809's but have never tried to use them. I have tried 811A's. Two stage amp. Same circuit as the TubelabSE. They work good. I use mosfet followers to supply all the grid current that any output tube can eat.

Click to expand...

Thanx Tubelab. That is an idea that hadn't crossed my mind. I'll consider that as a backup plan if I hit a dead end.

Here's the compulsory tubes shot.

Thanx muchly for the compliments Guys!

As of tonight I have it assembled and all finished on the exterior. I still need to wire it and stuff the insides, but it does look purty. The front is a beautiful chunk of tiger walnut that folows you as you change the angle of view. See where I put the pilot LED? Too bad this rare chunk had a crack in it but it was such an amazing piece that I just had to use it. Some of the most amazing grain pattern hardwoods are quite sensitive to drying fractures. The only filler I had on hand is mahogany and shows up. I can address this later. You probably wouldn't have noticed it if I hadn't said anything right?

Here are some pics. Like my sexy miniature speaker binding posts?

Second view of four.

Third view of four.

Last view of four.

I have discovered a pair of vintage high quality Hammond 937 interstage xfmers in my inventory that I had forgotten about. I am going to use them in this amp and will be able to drive the 809 triode now properly into A2 where it wants to be.

In addition to the power xfmer and two P-P output xfmers on top which you've seen, putting all the extra iron neded which is three iron choke transformers, two large interstage xfmers, a filament power xfmer, two massive (the size of Cambell's soup cans) polypropelene, foil and oil motor run caps for the output stage coupling, four big stud mounted power R's, smaller electrolytics and miscellaneous other components all underneath inside that small wooden box has been a real challenge but I have succeeded. The only way was to make a sub chassis that would hang below the main top plate and hold all these other components as a complete module package which could be lifted out of the wooden case and worked on from all angles. I think I deserve some sort of a prize for my accomplishment. I will let you judge when I have pictures of the naked assembly to show within days or less.

Here is how I built this amp.

Time to share the build sequence photos and construction details of my amp with the group.

As I had explained I was after good looks on this one and that forced me to use a box, the largest I could make with available walnut mill scraps from my firewood pile, which were on the small side. Conspiring towards a space problem was the fact that I did not have real SE output iron, so in addition to a pair of Hammond 1615, 15 watt, 5k to 4-8-16 P-P output xfmers I had to shoehorn in two foil and oil motor run caps the size of a cans of soup and also two plate chokes (vertical output xfmers hand selected for balanced L and salvaged from late 1940's vintage RCA television receiver chassis). This project was ALL about available parts and creativity!

The massive (and sexy don't you think?) 5 way speaker terminals are impedance selected on any of the 3 output transformer taps with a ceramic rotary switch with parallelled silver plated segments driven by a user adjustable knob next to the main power toggle switch located between the speaker terminals at the top rear of the chassis. Best match to the speaker system may be obtained by adjustment. Mismatch is much more damaging to the performance when only 3 watts are available. The gold plated RCA input jacks and adjoining screwdriver adjustment input pots are at the front of the chassis top between the input tubes and the power xfmer.

I also had to find room for two rather large Hammond 937 interstage xfmers in order to drive the 809's properly in A2. BTW I have G1's biased at ~+9 VDC and with a plate cathode voltage of 390 VDC I am running ~48 mA in each output tube and dissipating about 18.5 watts. The PD is 25 watts for these tubes.

Also hidden inside are the main PS choke (7 Hy's@140 mA), and a custom wound dual filament winding power xfmer for the two DC supplies for the 1619 heaters. The two 809 heaters are in parallel sharing the same DC filament supply which uses the 6.3 volt winding existant on the main power xfmer. I just did NOT have room for anoither filament xfmer as to separate the 809 heater supplies and successfully worked the schematic to accomodate this handicap.

DC for the red LED pilot iis obtained from the positive DC 809 grid bias supply employing a 3 watt, 8.2 volt xener circuit and the LED in series.

A humble amount of DC power for the computer PSU salvaged 4" cooling fan is also pulled from a voltage divider chain in the amplifier circuitry. The fan is mounted externallly under the amplifier's bottom plate, hence the need for tall footers. Might as well run the fan instead of heating a power R with this limited energy supply.

I had to be clever about power sources as my power xfmer was sized marginally (using available parts). I get B+ power for the driver 1619's by stacking them in series under the 809's. That way the 625 volt B+ at 120 mA available has enough power to run both 48 mA amp circuits and the 22 mA G1 bias supply circuit. The cooling fan is supplied through the 809's.

B+ supply is a 720 VAC (CT unused) winding feeding a hybrid bridge (two 1N5408) and a 5V4 indirectly heated cathode rectifier tube. It is like a 5AR4 but comes on faster in 6 seconds but still allowing adequate startup delay and IMO looks better with the tall ST shape. Choke input filter to a pair of 550uF@450 VDC caps in series with parallel EQ bleeders. This makes 620 VDC under load.

Originally configured as pentodes, try as I might I could not get a good match or sound from the input stages with my only I.T. xfmers available. I had ruler flay response from 100Hz to 100 KHz but there was a kink in the waveforem, no bass and it sounded very bad. Rewiring the drivers in pure triode mode solved these issues.

The circuit, my own, was entirely experimental and no part of it had been breadboard tested prior to what was a very effort intensive construction process over about a week and a half period. On first full power up upon completion there was no smoke, voltages were pretty close to intended and it made music but sounded like a Western Electric amplifier. "Number please?". Fortunately within just a few hours I had experimentally developed a NFB loop featuring frequency taper that provided about 4 dB global voltage FB which extended the useable bass spectrum and tamed the amplifier to the point that in the lab it was sounding very good and looking quite acceptable on the o-scope with sinewave testing. Useful output power was measured into an 8 ohm load at 3.1 watts RMS/ch. At this point it was put in the case and shipped to the living room where my best loudspeakers, the efficient Ultor-X design, have found an appreciative home. More on listening impressions later.

A star ground was religiously adopted which worked. The amplifier is dead silent. and with the built in line filter IEC power connector, is also immune to lamp dimmer buzz on the house mains, a real problem with some vintage tube amps like the Eico HF20's and Heathkits I've tried in my living room.

I will post the schematic when I have a chance to draw it into something comprehensible by others.

As far as construction, the usual stuff and the best looking parts got put on top of the 3/32" thick solid yellow brass top plate. In order to stuff a lot of parts inside I made a sub chassis that would hang about 1" below the main chassis plate. The tube socket wiring, various PS diodes, filament supply filter electrolytics and such all fill the gap between the two chassis. Things were so tight I had to use miniature input gain control pots and had to mount the three protection fuseholders in between the chassis, accessible only by removing the guts from the case. This is not hard to do as the entire guts pull out of the wooden case as a single mudule as you will see.

2nd pic of nine.

3rd pic of nine.