Audio Asylum Thread Printer
Get a view of an entire thread on one page
|For Sale Ads|
As you can see the results of the rebuild of the now beautiful H50 here
I figured it was also time to give an update of the results of the identical pair of HO50 which were rebuilt in 2019.
In that case I decided to leave them with their original 110V transformers, and buy in a superb USSR (Polish) made 230-110V convertor to power both of them, housing that in a nice little box, ordered in specially.
The transformer is rated at 0.95KW...and terminated in those gorgeous little Bulgin plugs - one for each channel,to make it easy.
The tops were blasted and powder coated, then some special transfers made to put the DB insignia back on them.
The bases were also blasted and powder coated, but warped a little with the force of the blast process. They only sit flat, once screwed back on the cases.
There's a reason to have the tops and bottoms all neatly screwed on.
You don't want your kids going anywhere near that exposed 500V internally!
Of course the mods to the original circuit are really numerous.
The -42V negative bias lines being derived via a stack of 5W zeners, lurking in the HT supply transformer CT return.
This was a tough trick to execute properly, as the decision to change the entire amp to rare matched pairs of genuine GEC KT8C also turned out to be a struggle.
(Basically this converts the 807 amp to NOS KT66, on an ancient B5 base, but an affordable price instead of the insane prices now worldwide for the octal base genuine GEC KT66.)
Of course the KT8C takes 30% more heater power than the 807, as it has a much fatter larger cathode. 1.27A instead of the 807.
Compare the 807 versions with the KT8C
How does this ancient perform, you may ask?
Well, it was directly compared against a fully modd'ed MO200,a Bogen DB110 and several other brand new "hi end" amps costing lots of money.
We also ran it into Vandersteen brand new speakers & Amphion.
The result was quite a surprise, being as they have an output power and dynamic range which exceeds most moderns we tried.
A 70yr old amplifier that works as well as this with balanced line input too?
I have to assume you mean the Stancor Williamson amplifier CH-2133 rated at 10W??
These were sold as kits with the A-8054 OPT, and the 807 screen grids tied to the anodes in an attempt to imitate the PX25 triode.
There were major problems with Williamson amps, because without a decent transformer they turn into great RF oscillators.
As the British knew, they turned to Partridge for a proper interleaved winding arrangement.
In reality the KT66 with higher screen grid ratings is better than the 807, which is why DTN W used it.
This triode curve tracer graph is for the KT8C, which is, an exact (war time) forerunner of the KT66,- a failed transmitter valve.
Stancor made a Stancor A-8072 "ultralinear version" which is an anachronism. 6L6/807s and the like with their aligned grids don't basically work with "ultralinear" because the screen grids have inadequate gain.
Probably would work a lot better with the EL37, forerunner of the EL34.
Apparently Stancor, UTC with their LS series and Peerless were about the only people capable of making decent interleaved OPTs.
The Bogen amp seems to be pretty impervious to HF oscillation,I suspect because the OPT is well made and because it's AB2 is HALF the usual A-A load. (ie. low impedance).
I never checked the KT8C to see if you could run the screen grids at KT66 voltages, but the only data I have is stating 300V MAX, although with G1 +10V they make 350m/a Ia at 100V, +20V quotes 430m/a.
RCA is quoting much higher at about 0.5A from the 807 at those voltages, but I can't see how it can do it with such a small cathode.
(KT8C has much higher perveance.)
I dismantled one and measured it.
Here it is:-
The cathodes are huge in the KT8C, but the wire failure rate particularly in the wires that went down into the (ceramic) base to my mind were totally unaacceptable.
Then one guy sold me one "so called" NOS which in fact was a complete pile of rubbish burnt out junk...
That put me off them for good.
Cathode designs took huge strides forward as TV horizontal scans and military needs demanded them.
My little mil spec pentodes designed in 1955 for nuclear bomber radios & missiles (& still made in 1970-73) have cathodes 30% larger than the 2nd World war 807...but a package 1/20th the size.
This is how a modified one looks underneath.
It's a little difficult to explain how to move from dropping resistors on the original amp, which is fundamentally unstable to a much better solution with modern zeners.
A shunt stabiliser is run off the main 500V HT via the original dropping resistor chain (which used to be across both the screen and anode supplies to ground).
The visible wire ended one under the chassis is run immediately at the full max of 0.040A in series with another one above chassis in a loctal base, where the original 5R4GY used to sit.
This NOS one was from the BBC from years ago...
As the anode supply draws down from its initial 600V+, (as the output valves warm up), and the AF amp valve heater warms up, the gas discharge valves pull back nicely to a more moderate 0.020A, and the new HT supply settles at 505V, a far more RIGID 500V than ever before despite having 32V drawn off it on the string of zeners in the CT return.
The 5R4 was moved from its original spot to next door instead of the 5Y3, 'cos of course it provides a much lower voltage drop for a screen supply than the 5Y3, and gives a warm up gentle start.
That was also fed with silicon diodes and the ones in the 5R4 put in parallel.
This is quite a smart mod.
The main Anode supply comes straight out of some silicon diodes into the swinging choke, then straight into the dropping resistor down into the gas stabilisers.
There's no non linear voltage drop thru a valve rectifier.
They act as surge limiters, by lighting up straight away, and also giving an immediate return current via the zener diodes in the CT of the main anode supply.
That brings the 32V bias supply up immediately well before we start the screen grid supply up.
So the gas discharge valves, derive the stabilised HT supply for the AF amp section (minus hum and noise).
Originally all that was derived via a dropping resistor from the screen supply.
Clearly as the screen/driver current varied, so did the supply to the AF amp.
807s and KT8 have aligned screen grids, so in reality the screen supply demand is not enormous.
Here's some comparisons of how the original amp fared at merely 15W (distortion started to rise rapidly), compared with how it now manages a really linear 60W after being sorted.
There were some horrible crossover distortions and instabilities taking place especially between 20-100hz...and nowhere near the 50W it was supposed to make.
The paraphrase invertor was clearly not happy, needing a balance pot to make it better.
Most people would of course have given up here, and said, "NAH its just a rubbish 1940s PA amp init, something about silken purses and sow's ears?"
Well, if you gave up then that you would be wrong.
Bogen have excellent output transformers wound themselves on Selectron cores, all that was really needed was balancing the phase invertor correctly, stabilising the power supply, and REDUCING the original amount of NFB which was overloading one side of the AF amp while giving no help to the other side which wasn't.
It's the main reason why people threw Bogen amps in the bin claiming they could never be any use for hifi.
So, take a look at this after well sorted.
60W square wave?
70W 1khz sine wave?
Who says AB2 amps based on the 807/KT66 are dirty with gross distortion?
This Bogen is an absolute gem and it's 70yrs old!
A ruler board frequency reponse dead flat from 30hz-15khz, with no sign of instability into any load?
The irony was not lost.
The shop where we tested it, didn't like the sound.
"too neutral" they said, "too clean"..another said.
What to say?
They loved the sound of the original 6V6 powered DB110, which was way less powerful, extremely coloured with a great big lump in response at 100-300hz typical of valve amps!
One of my favorite amps is the two chassis Stancor 807 mono amp with the triode output transformer. Enjoy!
Your amp is listed here:-
I have to say the IMD goes ballistic on the triode amp a soon as you exceed 8W, that's probably because it's cathode biased.
The ultralinear amp develops 2.5x more power but again as you exceed 25W the IMD rapidly exceeds 3%, again because despite the 400V HT, it's still cathode biased.
I don't imagine how that would sound with inefficient speakers and large transients.
I have to say the modified HO50 bogen amp sounds best with rock and jazz where it sounds a though there is no limit to the dynamic range.
Fixed bias 807s/KT8 running at 500V with DC coupling sound a lot louder than 60W!
Astonishingly for such an ancient amp, hum and noise is so low, I have to put my ear on the speaker driver to hear anything at all, despite it having really small smoothing caps.
I did arrive at the final (DUAL) NFB range by trial and error, and careful listening tests.
So far I have yet to see an implementation of dual loop NFB using the 140V centre tapped line output windings.ie. +/- 70V
It really DOES work.
I can believe Stancor sold 7000 of those Triode based williamson amp kits.
It was the least expensive way to hit the "quality" amp mantra in 1950.
I am keener on BIG amps, as they can feed big hungry low efficiency speakers.
I guess that's why I became so interested in turning Bogen monsters into true hifi amps.
I think our Russian friend Anatoliy has a lot of it right:
He especially holds the correct view, 6L6 oids sound no good until right on the limit of melting, which presumably is why Williamson gave up and connected them as low power triodes.
1940-50s Bogens are a lot rarer.
I mentioned this before, but I have this pair of HO125, which has been on hold, wondering how to tackle it, being as someone has already melted lots of bits in them.
807s had special anodes and a lot of other tweaks to make them stand up to intense abuse, which of course is exactly what I intend to give them!
Having ONE is already quite rare.
Having a pair is really a one off, and there are none in Europe.
(I have heard of one other person managed to acquire a pair in the USA.)
One was a rack mount, so it's missing the cage, and has been horribly bodged. Not sure what can be done to find another cage.
Here they are sitting on the floor of the store room:-
After careful checking and blowing fuses, it turned out, BOTH HT mains transformers had internal shorts.
This amp has seperate heater and HT transformers, but the heater transfos are OK, just need a small toroid to feed them 110V via a simple step down.
I have a surplus pair of 110V HO50 transformers, but they are simply not up to supplying so much current for a quad of 807s.
I have a toss up which way to rebuild them, but sure as anything now they will get a matched QUAD of 807s instead of the original heavily driven pairs.
Being as the original 807 bottle is huge, that inevitably means changing the sockets to brand new LOKTALs, adding an extra pair and drilling new holes with a purpose bought Q max punch.
The HO125 is the only one that is capable of being modified this way,- more space in the giant chassis!
I have already made up 2 matched quads of brand new NOS 5B254M (which actually cost almost nothing), each one being tested for current under working conditions in the old HO50 guitar mixer amp.
There were 2-3 which arced over inside from new, from a batch of 20+, which proves a point,- there's no point in buying sub standard Chinese or Russian modern stuff when even original industrial/mil spec UK made stuff has a high reject rate from new.
Here are shots of the incredible Bogen HO125 pair from underneath.
Beautiful tag board construction, making life easy for replacement of the various bits that have aged and died.
It's remarkably similar to the HO50, except for paralleled 6SN7 driver valves, a HUGE output transformer using the same but even larger cores used in the much later MO200, and of course anode HT of 835V on those top caps.
These are 70 years old remember.
Being as they are AB2, and suffer zero blocking distortion, so will give any 2020 KT150 amp from Upscale, Primaluna or Jadis such a serious run for its money, people will wonder what has happened in all that time.
They must have cost a lot in 1949, there's simply so much iron on there, which is why the amps weigh a ton!
Another funny thing.
People see the 90ohm, 70V and 140V taps on the back and give up!
People claim (of course wrongly) they can't be used for normal 8/16ohm loads, so they throw away an outstandingly good transformer.
Well, in those secondary windings they have 2 centre taps.
The com - 70v and 140V, of course the 70V is CT for 140V.
However, this is really smart, the 90 ohm winding (105V) is the CT for the 70-140V winding which gives guess what?
2 x 35V taps. ie 2 x 10 ohm windings for 125W.
Being as we now know the output transformer A-A load on the HO50 was an astonishly low 3.3K ohm, I measured the HO125 OPT, T254BU.
This came in at 5.2-5.3K.
I don't mind dropping the A-A load by connecting 8 ohm speakers on the 35V taps. That then gives a refected impedance of 4250ohms, - exactly half way between the Bogen MO100A and the quad 8417 M120,
(which also gives 120W,-I have a pair of those also).
In my opinion the 125W RMS figure is highly conservative because it only requires 800V RMS across the transformer.
Presumably, at which point the HT transformer ran out of current or the 807s themselves.
Basically we only need to swing 900V linearly across them with a nice strong power supply that doesn't sag like the original, and it will deliver 190W.
In my experience with the other Bogen H50s, the silicon diode fed amp had only 100V sag to 500V, (instead of 400v), easily made 70W clean instead of a very dirty old OEM 30W.
If that's the case I would put a bet on it, based on a 50W anode rating, they can deliver TWICE the original power at more like 250W....
We will see!
Being as the OPT are a full INCH bigger than those on the equally conservative rated MO200 (and the even smaller M120) - (with an octet of 8417s!) it's very clear the HO125 is the biggest beast, Bogen ever made, quite easily capable of close to 200W at right down to 20hz with likely very little saturation.
Here's one from a MO200 (same one as in MO100A).
With the top off?
Those come in at 4750 ohms A-A, which is why they make a good solid 110W RMS clean from 660V HT once the drive circuits are all sorted.
D Gillespie at Tronola has proven this.
That of course is right on the limit of what the Octal sockets will stand, as witnessed by many arcs and burnt out anode resistors on one of the pair of my Bogen MO200s. (Yes I have a pair of those fantastic amps too!)
The Bogen HO125 looks very promising because not only do the anode wires come out of the top of the huge drop thru transformers but the sensitive bits are hidden well below the chassis.
Here is what it's slightly smaller cousin looks like with the same number of transformers (!):-
They really don't come up for sale very often.
It's a sort of "detuned HO125"
They have a slightly lower supply of 765V,(still over the 750V thought to be the max in ICAS service for 807s).
It explains the lower power output of 100W, useing the same paralleled 6SN7 driver valves, to supply all that grid current but no bias stabiliser as in the HO125.
That is the main problem with the MO200s I have.
There is such tiny space between the high gain AF amp gain stages under chasis and the output transformer wires, it takes nothing to set the thing off into violent and self destructive oscillation!
Been there done that!
So, the idea is to convert them to matched quads of 5B254M (smaller anode caps) and find a power transformer that is compact enough and powerful enough to feed them.
Finding a drop thru that can produce 820-900V stable at 400VA after a swinging choke is proving no easy task, because it has to produce 1000V AC with minimal derating at full load.
A typical transformer manufacturer gave me these derating figures:-
On the 100VA - 10%
On the 200VA - 9% = 910V - after choke input filter 820V
On the 300VA - 3% = 970V - after choke input filter 870V
In this case the target figure for the new amp is 175W RMS clean at 1% THD.
Here are the diagrams for the JO100 and HO125 just to see the differences:-
Post a Message!
This post is made possible by the generous support of people like you and our sponsors: