required specs of an input step up transformer, into a 45, to allow it drive an 845

[rick57]

Hi

The idea is to do a 2-stage 845 SE, using a 45.

It has been done commercially by Electronluv (Ed Sawyer has one). There’s Ed says, is basically:

Input transformer (600: 15k???) >
45 (B+ about 200 v) >
1:1 5k IT >
45 output (900 v) >
(And parafeed output trans/ plate choke).

Ed has pointed out that the key to doing an 45-845 in two stages is (not just the driver tube alone) but getting enough gain, hence the use of a step-up input transformer. It would need to be a good quality step-up trannie (maybe custom) – eg James, Electraprint.

Making it a tad easier, 10 watts would do: the priority is hearing the sonics of the sweet 45, with 4-6 times the power, with a minimum of extra “layers”.

I believe long time 845er Grover Gardner DIYed one, but neither Ed nor Grover has a schematic . . hence my question

Recent 'underwhelm-ment' with some much touted SS amps has heightened my enthusiam.

So does anyone know the logic to calc the required specs of an input step up transformer, into a 45, to allow it drive the 845 to about 10 watts?

Thanks

[jeff mai]

Inside joke (sort of): Crowhurst!

[rick57]

Jeff or anyone

(Jokes aside, without encyclopaedic detail) is a summary or link possible

Cheers

[Ed Sawyer]

I looked up the schematic for these amps last night. They use mainly custom Electraprint iron with silver windings, here's the basic details:

input trans: 1k to 15k (600 to 10k would also work or 600 to 15k)

45 runs at about 200v B+, battery biased. DC filaments.

1:1 IT (5k nominal rating, hopefully something like 40-50h+ would be nice, or parafeed it if you want with a 150h+ plate choke)

845 runs at about 900v B+, 5k:8 ohm Parafeed output trans, EI 125 lams. Battery and 1k cathode resistor bias. AC or DC fils (DC for lowest noise). 200h+ plate choke @ 100ma. approx 1-3uf parafeed bypass cap.

That's pretty much it. Power supplies depend on what you are into. These run MV rectifiers in a dual LCLCLC supply setup. You can tweak op points based on where you want the B+ to run. LMK if you get any further with this and I can provide any other details if needed.

Duplicating the amps exactly would be difficult, as sometimes Josh's work differed from the schematic, plus all the iron was custom anyway. But it's a good start (the above info)

thanks
-ED

[jeff mai]

Originally posted by rick57

(Jokes aside, without encyclopaedic detail) is a summary or link possible

What I said was sort of inside, but not a joke. The material you have IS a summary.

[rick57]

Hi Ed

Thank you!

mainly custom Electraprint iron ~ ooh mucho $
Duplicating the amps exactly would be difficult
I see your point.

For those who don’t visit the Tube DIY Asylum, some alternatives *without input step up transformers emerge here:

http://www.audioasylum.com/forums/tu...es/108682.html
http://www.audioasylum.com/forums/tu...es/108614.html

Cheers

[jrdmedford]

Heya. I'm the Jim who did the V3,V4 and V5 low-voltage 845 amp schematics
that people seem to link to alot.

A two stage high power 845 amp is possible, however, there are tradeoffs to
how and where you get the voltage gain required to drive the 845's grid to
full power. Lets look at tradeoffs.

Option 1:
You could use 45 driver with a 1:2 stepup transformer. However, the transformer will now reflect the Miller Capacitance + Transformer stray capacitance towards the preamp and cable. This will force some extreme requirements on your preamp: You will need a very low output impedance preamp that not only has tons of current grunt but also an output level of +10V. There are solid state preamps
that do this job very well. You will surely be building ANOTHER tube preamp, or
tube line buffer, with stepdown line out transformer to get the drive and low impedance you need. On the other hand, a solid state differential line driver
(i.e. Analog SSM2142 or AD815), might sound better and be far cheaper.

Option 2:
You could use a single high gM tube driving a 1:2 interstage. Consider a
7788/E810F (gain=52, Rp around 700 ohms). Other tubes include the 6C45
or the EC8010/EC8020. You'd use the Lundahl LL1660/25mA in the 1:2 stepup
mode. This too will work, however, you are at the mercy of finding driver
tubes with matched gain. You will also be at the mercy of UHF oscillation and your circuit will need treatment to avoid the excitement of parasitic oscillations.

Option 3:
You could do what I did. Use a 1:1 interstage. Simulate "two stage" by using
a direct coupled driver of the interstage. Direct coupling eliminates a phase
shift which eliminate the problem with 3-stage amps - stability. If you chose
to go 3-stage, you can. This seems to have the fewest tradeoffs. High Z input stage, easily produced voltage gain with simple devices, stability.

In the end, keep the input transformer (1:1) regardless of what option you choose. Moving balanced signals around your house is so much easier than
single ended (no ground loops, etc). In addition, I like the galvanic isolation
that exists between my amp chassis, and my low-voltage, finger friendly shelf of source devices.

-- Jim

[rick57]

Hi Jim
Ah the man behind the circuits - thanks for your further information!

BTW, which of your three versions do you prefer?


Personally I don’t seek high power, *as long as dynamics aren’t compromised, 10 watts is *plenty.

Option 2:
Interesting, though 7788/ E810F/ 6C45/ EC8010/EC8020 don’t appeal greatly. I would prefer DHT or at least DHP (eg the 47) . .

Option 3 might be good if Option 1 can’t be made to work well with what I have ~

Option 1:
A SS preamp is not heresy, if you think it good enough to not introduce grain etc. Where for a * Good Analog SSM2142 or AD815 circuit (Analog’s site??)

as I have a potentially suitable preamp: the Mapletree Magic 5 -

For a 45 with a 1:2 step-up, a preamp is needed with:
· very low output impedance
· has tons of current grunt
· an output level of +10 V.

* Output Impedance
I will be using an autoformer (TVC). Their output impedance varies with volume, but in most listening will only be (I only just learnt this, hope I got it right)
e.g. 10% volume: 1 volt from the CDP; TVC out of 0.1 volt:
Output impedance = 100 R / ((1/0.1) squared)
= 100/ (10*10)
= 1 R
So at 10% volume, the output impedance is only 1 R.
At 50% volume (i.e. loud!), 1 volt, TVC out of 0.5 volt
Output impedance = 100R/ ((1/0.5) squared)
= 100 /4
= 25 R
As the most likely drivers are 98 dB, output impedances of 0.5 – 10R should be low enough?

* Output Level
The Magic 5 into a load of 20 K or more, outputs well over 10 V rms.

* Grunt
How many mA in a ton?
Into a power amp, the Magic 5 current delivered depends on the input resistance of the power amp. At 100 K load and 4 V peak (around 3 V rms), current would be 0.04 mA peak (around 0.1 mW)
I have to check what would happen if bypassing the pre’s volume control with a TVC.

Does Option 1 with that pre leave anything not addressed

Thanks

[phn]

I would go with the 45. The output tubes will never sound better than the drivers.

I think 20k (as in 600:20k) will work as well. The 20k the gurus here will need to confirm.

I would use interstage transformers as well.

Chance is that guys like Ken Shindo have been wrong about everything for 30 yrs. You can believe that if you want.

There are only one of two reasons Shindo gear sounds the way it does. Magic tubes. Or input, interstage and, in case of preamps, output transformers.

[rick57]

> I would go with the 45. The output tubes will never sound better than the drivers.

I'm a babe in the world of tubes, but tend to agree

Chance is that guys like Ken Shindo have been wrong about everything for 30 yrs. You can believe that if you want.

I haven't heard of Ken Shindo:

> only one of two reasons Shindo gear sounds the way it does

which I take it is really good in some or several ways . . ?

What is his site?

Cheers