6550/KT88 SET amp aka "Uber Beam Machine"

JoshK · « **Reply #100 on:** 13 Dec 2007, 12:11 am »

I totally forgot about that. Those datasheets are indeed scans/photo's whatever. Its a product of them being from the 50's, 60's at the latest.

If I can help, let me know. I can type up some of the data from the datasheets here, as there isn't too much to write. However, the charts I don't know how to help there.

I appreciate your fervor. Its probably got to take you a bit longer than anyone else given the lengths you have to go through just to read the data.

jrebman · « **Reply #101 on:** 13 Dec 2007, 12:31 am »

Yeah, and that's probably why I sound a bit dense and slow at times -- access to graphs, charts, and instrumentation is the thing that gets in my way, especially being a very visually oriented learner.

I will say that this project is giving me real motivation to figure out my tactile graphical printing software and the printer it works with.

Which is also giving me good information to feed back to the software vendor on how to improve it's capabilities. This stuff works, albeit quite crudely at the moment (for these purposes anyway), and still requires a good deal of assistance to use effectively.

Note that I'm not complaining -- this is all learning and problem solving, which is where an engineer/computer geek like me is happiest.

I'll soon also have the help of Dave C -- an engineering student and who lives a couple miles from me, and another guy who has a science/technical background, which is something that I just haven't had in the past.

I've got my "eye" on the prize and I'll get there -- just not as fast as I would have 20 years ago.

And for the record... it's fun!

-- Jim

jrebman · « **Reply #102 on:** 13 Dec 2007, 12:33 am »

Thank you also for the offer to help withe data translation. If I need that, I'll let you know, but I think I can manage with the resources at hand.

-- Jim

JoshK · « **Reply #103 on:** 15 Dec 2007, 12:29 am »

Just thought I'd mention, the $6300 award winning 845 SET amp named Monaco from Ultra-fi uses none other than....drum roll please.....James 2123H!

http://ultrafitimes.com/
http://www.positive-feedback.com/Issue32/moore_845.htm
http://www.positive-feedback.com/Issue34/awards.htm

They even mention something about it being made on the same assembly line as Tamura.

JoshK · « **Reply #104 on:** 17 Dec 2007, 04:51 am »

OK, well I measured my PTx, but since they aren't obtainable anymore, someone else is going to have to tell me their measurements of their PTx (simple resistance measurement of primary and secondary) if they want me to help them with their power supply. I'll share my supply and what those values were in case anyone wants to know, but after I have actually built it and validated it measures in reality the same as it measures in sim.

If everyone else is using George's PCBs then you probably don't need a PSU schematic, since you can use the one suggested in the kit and built into the board (although he gives options for chokes offboard, I suggest you opt for them).

The holiday rave and cleaning for my sister-in-law to visit tomorrow meant I didn't get anything done on the amp this weekend. I intend to built it up on a piece of plywood first. This way I can change out the front end and test a few other ideas before I commit to a design and layout of that design. This is always a good step as I see it.

Good news is that a reasonably simple PSU does the trick if values are chosen well. Choke input yields a nice sinusoidal ripple too.

nodiak · « **Reply #105 on:** 18 Dec 2007, 11:12 pm »

just keeping the thread warm...bein the holidays n all am not expecting to purchase parts yet. Have been listening to an Eddie Vaughn EL84/5r4/6922 with Tamura F-475's. Smooooth in triode, and razor sharp in UL. Should help suggest some preferences for the more powerful UBM to come.
Rain here in the Pacific NW, heard there's various storms in NE usa, how's it in central Rockies and Peru?

jrebman · « **Reply #106 on:** 18 Dec 2007, 11:31 pm »

Don,

We had a little snow last week, but it's been in the 40s for the past few days, so pretty nice overall.

We are making good progress here on the trio of amps -- sent in a big digikey order yesterday, and one of the other guys is sending in a partsconnexion order today or tomorrow.

I did all the calculations for operating points and power supply, and spec'ed the PT, and settled on a 5k OPT impedance. It's all really a juggling act of compromises, but I think we found a good place, and interestingly enough, we'll be getting more power from an EL34 than a 6550, but we'll be able to use all the tube types --EL34, Kt-88/6550/6l6 and get good power with minimal distortion and better damping factor across the board than we would have if we optimized for one tube type.

As things stand now, it looks like the chassis will be the bulk of the work.

We're havin' fun!

-- Jim

nodiak · « **Reply #107 on:** 19 Dec 2007, 04:53 am »

Sounds great Jim, very nice to have a few people working on amps in person for help and comparisons. Did you guys decide on transformers? I'm still looking around, am open to ideas.

DaveC113 · « **Reply #108 on:** 19 Dec 2007, 05:18 am »

nodiak, we are leaning toward James OPTs and settled on the Allied PT mentioned on the SimpleSE site thats supposed to be the same as Hammond for less $, but Jim is probably going with the smaller 20W Hashimotos. We are deciding on what type of resistors and caps to use, once we figure it all out we'll share.

On another topic, I want to replace the input potentiometer and the 100r / 220k voltage divider circuit between the pot and the 12at7 with a single voltage divider.

schematic:

http://www.tubelab.com/SimpleSE_schematic.htm

If the wiper of a 75k pot is set halfway (6 dB attenuation), that would be equivilant to two 37.5k resistors as the input pot. Looking at the pot and the 1st voltage divider, it looks like the signal path resistors are in series, and the ground legs are in parallel.

Adding the resistors in the signal path gives 37.6k, and combining the ground legs gives 32.1k. Does it make sense to replace the pot and 100r / 220k resistors with ~38k and ~32k resistors?

Also, my preamp has 29 dB gain, ideal amplifier input impedence is 50-100k.

Thanks,
Dave

DaveC113 · « **Reply #109 on:** 19 Dec 2007, 08:46 pm »

This should make it easier to see what I'm trying to do:

Dave

JoshK · « **Reply #110 on:** 19 Dec 2007, 09:18 pm »

I'd leave the 100r resistor alone. This is the "grid stopper" resistor, meant to keep the tube from oscillating among other things. There is nothing sacred about the 100r value but it doesn't need to be changed to do what you want to do since it isn't part of the voltage division.

My $.02 if I may. I wouldn't use so much gain in the preamp and then attenuate at the amp's input just to up the gain again with a high mu tube such as the 12at7. If your preamp has that high of gain, may I suggest you change the 12at7 to a 12b4a? 12b4a is a fantastic little tube. It is a single triode, so you need one per channel, so I am not sure if that will work with George's boards. If not, then how about the ECC99 from JJ? It has a mu of 17 instead of 70.

George's choice of the 12at7 was largely due to the need for the gain since he intended it to be driven directly from the cd player. If you have a lot of gain ahead of the driver stage, don't throw it away by shunting it to ground (also increases noise) and then regain again. Use a lower mu tube for the driver and take advantage of the gain from your preamp.

DaveC113 · « **Reply #111 on:** 19 Dec 2007, 10:06 pm »

Josh, thanks, that sounds great. Does the ECC99 use the same tube socket as the 12at7? I am using the SimpleSE pcb and only have room for one driver tube. My preamp does have a lot of gain, it uses 4 6922s in a mu-follower circuit. I'd like to keep it, it can swing 44V rms which would work great to drive a Firstwatt f4, my next amp project...

If I use the ECC99 tube, would I have to change anything else in the circuit? My post 3 back has a link to the SimpleSE schematic.

For the input, do you think I should use the 100r and 220k resistors just how they are and eliminate the pot without any compensation? (assuming I use the ECC99) How will this effect the input impedence of the amp as I am shooting for 50-100k...

Thanks,
Dave

JoshK · « **Reply #112 on:** 19 Dec 2007, 10:22 pm »

The ECC99 has the same socket/pinout as the 12AT7, it uses 12V heaters just as the 12AT7. It is similar to the 5687 in other parameters but the 5687 has a different pinout and uses 6V heaters. The ECC99 is going to draw more heater current than the 12AT7 but I'd have to imagine you should be ok, even with a cheap transformer, since it doesn't draw a ton of heater current, especially with only one tube. The bias points though will be different for the ECC99 than the 12AT7. This will require that you change the resistors in the anode and cathode sections. Not hard though and we can help.

For the input, my guess would be to eliminate the pot, leave 100r as 100r and change the 220k to 100k. That would give basically a 100K input impedance. You might confirm this with George but this sounds about right to me.

DaveC113 · « **Reply #113 on:** 19 Dec 2007, 10:54 pm »

OK, that sounds like a good solution... the only missing piece is the new anode and cathode resistor values. Just to be sure we're on the same page, were talking about R10/R20 as the cathode resistors, and R13/R23 as the anode resistors in the SimpleSE schematic?

If you could help me with the resistor values that would be great... We're planning on ordering from pcx before the sale (15% off film caps) ends on friday.

Dave

Dmason · « **Reply #114 on:** 19 Dec 2007, 10:56 pm »

Can you publish or post or email or PM a shopping list?

JoshK · « **Reply #115 on:** 20 Dec 2007, 12:40 am »

Dave,

At first eyeball it looks like you don't need to change anything. He mentions running the 12AT7 at 10mA and shows a 220r cathode resistor. I have a couple of circuits where the ECC99 is used and it too is run at 10mA with a cathode resistor of ~200r. The IXYS chip presents a really high load for the tube, which is fine as is. The only question I have which I can't see easily from his schematic is how much voltage does he have on the anode of the 12AT7. If it is similar to the schematics I have, then I'd go ahead with a simple substitution with no changes and then tweak as desired.

DaveC113 · « **Reply #116 on:** 20 Dec 2007, 01:04 am »

Ok, I'll give it a try. That means I can use a 12at7 too, but it sounds like the ecc99 is a better match.

Thanks,
Dave

JoshK · « **Reply #117 on:** 20 Dec 2007, 01:59 am »

yep, sounds like it.

DaveC113 · « **Reply #118 on:** 23 Dec 2007, 11:05 pm »

Here's the resistor list we (by we, I mean jrebman) came up with.

I made the chart, so forgive the terminology, I know its probably not whats normally used. Also, we are using a choke in place of R1.

JoshK · « **Reply #119 on:** 24 Dec 2007, 01:47 am »

very nice, thanks for sharing. I mostly stuffed my aikido board today and cut a piece of plywood for building up the "bread-board" version. I am first going to build the Aikido version, then I'll play with George's.

Some common terminology, if it helps...

your "ground leg" is commonly called "grid leak" (R21, R11, R26, R16)
your "grid" is commonly called "grid stopper" (R22, R12, R25, R15)
your "cathode to ground" is commonly called "cathode resistor" (R27, R17) you used this terminology for the driver

The 100R resistor to UL connection, I don't know what the common term is for that. I might call it the triode strapping resistor.

Anyway, I still understood what you meant and I suspect others would too, just thought you might like to know what they are called.