Using Sweep Tubes for an Amp Build

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
HI,
I have a bunch of 21JS6 Sweep tubes and would like to build a bass amp with them. These are 28 watt dissipation tubes, very similar to the 6LQ6s used in the McIntosh MI350. I will probably be using then in PPP (4 tubes) for 100+ watts output.
I'm posting here and not the instrument forum because I think more people here will be familiar will sweep tubes than over there. Hopefully Tubelab will see this....
Anyway, I have a few questions. I am familiar will the old Ampeg SVT that used 6146 tubes and the failures that those tubes caused in the SVT.

I will be using 180 volts on the screens so they won't be over-dissipated.

1) What is the highest plate voltage I can use with these in an audio amp application? I was thinking 600-850 volts.

2) Do sweep tubes reliably work when overdriven? Or do they melt down like the 6146s in the SVT?

3) I was planning on using similar output transformer loads as what the 6550 or EL34 would use. IE. 1900 ohms PP. Anybody here have experience with this?

Thanks to all for any input...
Daniel
 
1. 600 volts is fine, that is under full power, so idle will be around 700volts.
2. if the tv sets are to be used as basis, i will say they are..
3. i believe you can use those traffos for opt's..

search for postings by tubelab, he tortures these tubes for breakfast...
i am sure he will chime in moemntarily, this are types of thread he finds hard to stay away from, maybe smoking amp too...
 
Anyway, I have a few questions. I am familiar will the old Ampeg SVT that used 6146 tubes and the failures that those tubes caused in the SVT.

Not too familiar with this amp, though it would seem the failures were caused by abusing the screen grid. The 6146 likes a lowish screen voltage, much like the 807 and other 6L6-oids.

I will be using 180 volts on the screens so they won't be over-dissipated.

The spec sheet for the 6LQ6 calls for a V2K= 125V. I would stick with that, especially for a guitar amp.

1) What is the highest plate voltage I can use with these in an audio amp application? I was thinking 600-850 volts.

Way too high. These TV HD finals don't need more than 350VDC as they were designed to pull bigamps at low Vpk's. Maybe 600V for a Class B or Class C RF amp, but too much for an audio final. That'll have it well into Class AB territory.

2) Do sweep tubes reliably work when overdriven? Or do they melt down like the 6146s in the SVT?

I'd say not, if you don't abuse the screen. I'd keep that screen voltage down in that case. These TV HD types weren't designed with Class *2 in mind, and it's not really necessary, given the large cathodes and peak cathode currents that were obtained without taking Vgk positive.

The only possible problem is that these types can swing Vpk very low, and that could cause excessive screen currents. That's something to keep an eye on: not pushing the finals so hard the screens start glowing. That could lead to poofage. It's not so much of a problem with a reproduction amp, but the way some guitar players play, it comes closer to "brick on the key". Another good reason to stick with the screen voltage spec, since that's what TV HD duty is: max RMS power continuously with one consistent waveform, and just about the hardest condition possible.

3) I was planning on using similar output transformer loads as what the 6550 or EL34 would use. IE. 1900 ohms PP. Anybody here have experience with this?

Thanks to all for any input...
Daniel

That's one of the good features of these types as audio finals: you don't need a large RL (P-2-P). That makes for better OPTs. I did a project with TV HD finals, and a stock 4K4 (P-2-P) OPT was an excellent match. That was a music reproduction amp designed for low distortion, though.
 
I am familiar will the old Ampeg SVT.......Not too familiar with this amp

Brief history of SVT:

Ampeg brought out the first big watt bass guitar amp in the late 60's. It used 6 or 8 (don't remember) 6146 output tubes for 300 watts (rated) output. The standard speaker cabinet used 8 ten inch speakers. There were some famous flaming failures in this amp. There was the Rolling Stones concert at Altamont where the amp burst into flames, and a few others. The reasons were two fold, yet Ampeg released a statement that the Stones failure was caused by operating UK spec (230 volt) amps on US (115 volt) power. It would not even work on 115 volts!

The 6146's were fed about 350 volts on the screen grid which is far more than they can eat. A tube arc or runaway would ensue taking out the tubes, power supply, power transformer, or OPT. Usually all of the above.

The standard speakers could not eat all of the power the amp made (over 400 watts at clip) and one would eventually go open. Then another....then the entire cabinet fails to an open sending the OPT up in flames. Often the user connected the SVT head up to a different manufacturers 2 X 15 inch cab, with the same results.

Ampeg redesigned the SVT to use 6550's and the tube related failures ceased. The 6146 failures gave that tube a bad name in the audio world, and many people shun it still today. Keep the screen at 150 volts and it will live forever.

Answers:

1) I have an old HP regulated power supply that goes to 650 Volts at up to 1.7 Amps. I have tested dozens of different sweep tubes and have not found a tube that has a problem at 650 volts. I would test higher if I had more power supply!

Sweep tubes can eat kilovolts.....your OPT probably can not. In normal operation with a resistive load the OPT will see twice the B+ voltage. In an overdriven guitar amp with a real speaker load it can see more. The insulation rating of the OPT will impose an upper limit on B+ voltage.

There is a certain amount of bias current needed to eliminate crossover distortion. It varies from tube to tube, but ranges from 15 to 40 mA. The bias current and your B+ voltage will determine the idle dissipation of the output tubes. For good tube life, keep this to 60 percent of the dissipation spec. I have found that this imposes about a 600 volt limit on many tubes. The 6HJ5 is rated for 24 watts. It needs about 30 mA to clean up the crossover, so I can run 600 volts to dissipate 18 watts.

2) A sweep tube in a TV set is switched from cutoff to full conduction about 16 thousand times a second, every day for several hours, for a year or two.....it's idling in a guitar amp. In order to keep it alive two specs MUST be respected.

The screen voltage....The tube will see full saturation when overdriven. The maximum current in a pentode is set by the load impedance and the screen voltage. In a TV set there was usually an unbypassed screen resistor to set the peak cathode current. This causes distortion in an audio amp, so we use a regulated supply. Most sweep tubes work well at 150 volts, some need a bit more, but I have never needed more than 175 volts.

The peak cathode current....Every sweep tube has a peak cathode current spec. Abuse this, and sparks will fly....inside the tube! Some are quite high, 1.4 Amps for the 6LW6. This is set by the load impedance and the screen voltage. 150 volts with zero G1 volts is enough to violate the peak current spec on most tubes, so we set the load impedance for max current. A sweep tube saturates to about 50 volts with G1 at zero, so AB2 (positive G1) is not needed or wanted.

3) Most sweep tubes are limited by peak cathode current. This is set by the load impedance. A 1900 ohm OPT is 475 ohms per side with a pure restive matched load. Assuming a 50 volt saturation voltage, the OPT will see the B+ minus 50 volts when the tubes are driven to full saturation. With a 600 volt supply, this is 550 volts across 475 ohms, or 1.16 amps. With 4 tubes, its 580 mA per tube. Most sweep tubes can handle this load, IF we were driving resistors. An overdriven guitar speaker is anything but a resistive load. Maximum tube dissipation will become an issue with low load impedances. Worst case comes at 1/3 to 2/3 of max power depending on idle current. Test any extreme case amp in a dark room with continuous sine waves for plate glow. 4 6JS6's on 1900 ohms at 150 watts should be fine. You should be able to get there with about 500 volts.

I have limited experience with the 'JS6 since I don't have many. The US tubes evolved over the years, and the later versions had extra heat radiating fins welded to the plates. Japanese and Korean tubes did not. My experience has shown that these tubes can handle less abuse than the later vintage US tubes.
 
The 6146 failures gave that tube a bad name in the audio world, and many people shun it still today.

Original SVTs used either 6146 or 8298 tubes. I don't know if they were ever that much a revered "audio tube" but Ampeg most likely chose to use them because they had proven to be rugged in the service of New York's taxi cars.

Well, service in New York taxi radio wasn't obviously analogous to often overdriving the tubes in a circuit, which was largely derived from a cutting lathe amplifier and far exceeded the safe screen grid voltage limits. So yes, they were a failure in comparison to later 6550s.
 
Original SVTs used either 6146 or 8298 tubes. I don't know if they were ever that much a revered "audio tube" but Ampeg most likely chose to use them because they had proven to be rugged in the service of New York's taxi cars.

The spec sheet does mention audio usage for AM plate mod duty. No mention as to how these sound, since fidelity wasn't a big design consideration anyway. It also wouldn't call for the type of over drive you see all the time in guitar amp use, as that would cause "splatter" on the ham bands, and that's something to avoid.

RF duty is a good deal less demanding than audio. RF loads are almost always pure resistive, or very close to it, unless you mistune badly to cause a high SWR. Most RF finals have a "tune" mode that either cuts off the screen voltage, or at least greatly reduces it, so that you can tune up and load up before going full power without the risk of poofage.

300W from a quad was leaving no safety margin, so it's no wonder they blew, especially with such severe screen over voltage. The specs don't call for a V2K > 250V
 
Here in OZ local manufacturers used to use sweep tubes for Guitar Amps quite a lot as they were cheaper and more available than the 6L6, EL34 etc.
Goldentone brand was typical.
Pictures and schematics here:
Goldentone
I have just finished restoring a 1965 Goldentone 60W Bassmaster. 2 off 6DQ6 for 60 Watts out.
6D6Q and 6CM5 were the 2 common sweep tubes used in OZ so they were plentiful.

It is worth looking at the power supply schematics. See how they derive a 1/2 B+ supply for the screen from the centre tap of the high voltage winding.
This schematic is typical:
http://www.ozvalveamps.org/goldentone/1757bassmaster.jpg

This arrangement was used in the local electronics magazine guitar amp the Playmaster 117 too:
http://www.ozvalveamps.org/playmaster/pm117cct60w2.gif
It is worth fitting a reverse biased power diode across the screen supply cap, apparantly when in standby, it was possible for that point to develop a small -ve voltage. This was noted in their errata.
http://www.ozvalveamps.org/playmaster/pm116-7errata.gif

This amp has serious "bark", I love it, so much so that I have just purchased a 2nd hand a pair of Rola/Plessy 12UEG50 which were the original fit speakers to put it back to absolute original.

If designing one of these from scratch I would substitute a differential splitter for the Cathodyne (also called concertina or split load) splitter for better overdrive performance. Later models of the Goldentone did this.

It is also worth a browse through the sticky thread here:
http://www.diyaudio.com/forums/tubes-valves/211254-those-magnificent-television-tubes.html

Cheers,
Ian
 
Last edited:
This amp has serious "bark", I love it

I had a guitar preamp, two different ones actually, plugged into Pete Millett's big red board. I think the term "scream" fits it best......maybe "this one goes to eleven."

Pete designed a conservative 18 WPC sweep tube amp. I bought a board and pushed it to 250 WPC using 35LR6's. I also paralleled both channels through a single OPT for 525 watts! I never dared to play my guitar through it, or even connect my speakers for fear of setting the speakers on fire!

I dialed the amp back to 125 WPC and published the design in the "engineers amp" thread. Several people built copies of this design. I used a 6HJ5 tube which is a 24 watt tube and requires a small modification to the board. I believe that a 6JS6 will drop in to that board with no mods, but I don't remember for sure. Look through the thread for a list of compatible tubes.

I built 3 of the 125 WPC versions, one of which got some serious testing with an ADA MP-1 guitar preamp, and with a guitar preamp of my own design. Both channels were fed in parallel, with a single 8 ohm speaker on each channel. A 250 watt guitar amp is LOUD, and does qualify for "eleven."

Long thread, but good learning for sweep tube use in audio amps:

http://www.diyaudio.com/forums/tube...power-amp-design.html?highlight=engineers+amp
 
Hi Guys

The SVT eats tubes because the screen-stops are a joke at 22R each. Replacing them with 1k-5W and removing the parallel diodes lets a single set of tubes last a life time. Did this on a lot of amps and many other techs have followed suit, all to good effect. This works with sweep tubes, too.

The plate cap on the sweep tube gives it its amazing high-voltage capability. The high-output cathode gives it its amazing pulse capability. Some tubes have graphs showing pulses of over an ampere or two, but at a very low duty cycle.

The Traynor Super Custom Special used four 6KG6s to produce 250W using a relatively low Va=560V, with Vs=140V. They settled at this low voltage for reliability issues. After all, a guitar/bass amp is a piece of portable equipment that will be bounced and bumped a lot. For something that only you will move or that will never move, you can use higher voltages with appropriately increased Raa for the OT.

Have fun
 
Hi Guys

The SVT eats tubes because the screen-stops are a joke at 22R each. Replacing them with 1k-5W and removing the parallel diodes lets a single set of tubes last a life time. Did this on a lot of amps and many other techs have followed suit, all to good effect. This works with sweep tubes, too.

The plate cap on the sweep tube gives it its amazing high-voltage capability. The high-output cathode gives it its amazing pulse capability. Some tubes have graphs showing pulses of over an ampere or two, but at a very low duty cycle.

The Traynor Super Custom Special used four 6KG6s to produce 250W using a relatively low Va=560V, with Vs=140V. They settled at this low voltage for reliability issues. After all, a guitar/bass amp is a piece of portable equipment that will be bounced and bumped a lot. For something that only you will move or that will never move, you can use higher voltages with appropriately increased Raa for the OT.

Have fun

I had a bass player friend years ago that had 3 original SVTs in need of major repairs. I ended up getting them all up to spec, and for payment he gave me one of the SVTs; he kept 2, I got 1. They are great amps but I found them to be temperamental. Kind of like a super high performance car. When everything was good, they worked super; but very prone to having issues. Kind of like a high maintenance girlfriend. :)
Anyway, by chance, do you know what the plate to plate primary impedance of the output transformer of the Traynor Super Custom Special is? That would help me a lot with the amp I plan on building.
Thanks, Daniel
 
Hi Guys

I thought the goal was to build a 100W amp or so? That can be done using standard circuits, voltages and available off-the-shelf OTs. You can basically take any standard circuit and just add the low screen supply required by the sweep tubes, then set the bias range accordingly.

For maximum power from the tubes, you might run into the need for a custom OT. This is not really a problem inasmuch as there are lots of good places to get these - avoid MM as they are crazy hypemasters.

As for the Super Custom Special OT, I don't have specific data, but from the operating environment it is easy to calculate that Raa must be <2k. This assumes just 60V across the tube at peak current and an entirely stiff supply. The first assumption is likely close, the second is not. So, Raa-max < 1k5 to allow at least 10% drop of Va.

I can't say I was impressed with SVT as implemented. The power amp on its own has so many bandaids to make it stable that it has seriously rolled off treble response. It is definitely "an approach", but not really the best one for versatile playing. It was however appropriate for the context of when it was built, who built it, and who it was aimed at. TUT3 gives the details. otherwise, the circuit is "cool" from a design standpoint.

There was nothing inherently weak about using 6146s initially, as it was a sturdy US-made tube at the time and anything else that would work reliably came from the UK and was too expensive. Again, the reliability issues were related to too-low screen-stops.

As for screen-stops, there is no good reason for any amp to be built with too-low values, especially since 1995 when TUT came out. Large manufactures are slowly coming around to the sensible way to deal with it, since enough time has passed that it may not appear like they are following my advice. proper screen-stops are even more important in cathode-biased amps where the tube is already under high stress.

Have fun
 
Hi again,
My plan was for 100+ watts. I said this, speaking conservatively... In the back of my mind I knew it would be closer to 200 watts. My original query was to get basic operating parameters for sweep tubes used as audio output tubes. My questions were answered by everybody here; thank you all for the info.

My plan is to use an old Hammond output (1650T) rated for 1900 ohms primary. I figure it's close enough in spec. My B+ will be around 575 volts, screen supply will be around 160 volts, taken from the bottom half of an isolation transformer voltage doubler. The full voltage from the doubler will feed the preamp / phase splitter tubes. These power supplies will be using lots of filter capacitance, so they should be fairly "stiff" supplies.
I now wish I had bought a couple of those 400 watt Plitron surplus transformers that George has talked about in the past. They would have been great for 6 of the 21JS6 tubes in PPP, but how big does the amp need to be? In this realm, amps start getting impractical quickly.

One more thing... I totally agree with you about MM. Total hype!!! No way worth the price they're asking for. I love how they go on and on about the special MOJO needed to replicate original output transformer sound; where all that, is just a cheaply wound transformer with no interleaving between windings! Funny.
 
Last edited:
I use stoppers on both grids on all of my conventional amps.

G1 gets a 1K to 5.1K carbon comp on any class A1 or AB1 amp. If G1 is driven positive (not usually needed on sweep tubes) the stopper must be smaller to avoid distortion. In some cases it must be zero....a wire with 4 or 5 ferrite beads, but these are usually low Mu low Gm tubes where the grid is driven positive enough to draw considerable current. The 845 comes to mind.

G2 gets a stopper resistor, but the value and power rating depends on the tube and application. The typical sweep tube with a lot of B+ draws very little screen current except for the instant when the plate voltage goes to near zero. I have used a 510 ohm 1/2 watt resistor in a HiFi amp and everything worked great even when I dimed the volume control for a couple hours on a 125 WPC amp, however the resistor went up in flames the first time I played the guitar through the same amp. I settled on a 150 ohm 2 watt resistor in that amp.

I have built several screen driven amps using sweep tubes including the 6LW6. You can't use a screen stopper in this case. The Gm of the screen is usually low enough to avoid major oscillations, unless you tie the source of a mosfet directly to the screen....add several ferrite beads. Note that screen drive is not a good option for an amp that's driven to clipping often. You will blow the screen grid!

I have built some G1 driven 'LW6 guitar amps including a really big one about 20 years ago, but I don't remember the exact size of the screen stoppers. The resistance needs to be high enough to limit screen dissipation during square wave operation, but low enough to avoid loss of power and smoking screen resistors.

As with the EL34 the value of the screen resistor will determine the "tone" of the distortion created when the amp is cranked. Stick a scope probe on the screen (after the stopper) and watch what happens during hard clipping.

If you have a dual trace scope that can subtract one channel from the other, put a probe on each side of the stopper and observe the voltage across the stopper. It will be near zero until the amp clips, then it takes off, dropping the screen voltage and reducing the peak power output.
 
One more question, George...
Did you use grid stopper resistors to the LW6 tubes in your setup?
What about screen stoppers, or did you just feed 150 volts directly to the screens?
I guess that was two questions...
Thanks, Daniel

In my designs, grid stoppers aren't necessary since the feed is a Lo-Z grid driver. It's different where the source is Hi-Z. That's where you're most likely to see parasitics in both small signal and large signal stages.

Screen stoppers is a whole 'nother story, and that depends on the finals involved. 807s and other 6L6-oids, and some HD and RF types like to make snivets. In cases where you're getting snivets, usually 470R -- 1K5 will prevent that. Sometimes the spec sheet will specify a small, fixed, positive voltage for suppressors/beam formers.

When doing a project with 6BQ6GAs, there were no snivets or other instabilities without screen stoppers, so I added 560R/0.5W C-comp screen stops just in case.

I also like plate stoppers, especially when using the plate top cap types since the plate leads are longer, and can't be twisted to cut down the parasitic inductance. These plate stoppers are coils ( N= 10; #18; 7/16" ID ) connect right at the plate cap connector. Parallel with a 100R/2W C-comp conveniently slipped inside the coil to de-Q it so's it doesn't ring at its natural frequency. Four 470R/0.5W C-comps can be paralleled in case you can't find the 100R/2W C-comps.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.