Thank you all for the advice and considerations on the measurements of the output transformers.
Reading the various posts it seems to me that the best solution is to create a final amplifier, with for example two EL84 final tubes with terminals to connect the primary of the transformer under test. On the secondary of the transformer under test I remain of the idea of using a 20 ohm rheostat, even if I always have the doubt that the inductance of this rheostat could false the measurements. I was wondering if we should connect a capacitor in parallel to the rheostat to compensate for its inductance.

-- Cheers
??? Attilio

Rebonjour a tous,

The long posts raised my curiosity: WWhat about an accurate LCR bridge measurement ofd loads and speakers?

My Wayne Kerr 3245 precision inductance meter was just reworked and in cal, set at 1 kHz 1V rms.

As a power electronics designer I had boxes of load resistors.
I selected various types of 10 Ohm 150...250W resistors and rheostats.
I had not found a non inductive WW R nor Gol-Bar graphite rod R yet. I would assume the Globar L is just the lead inductance so some uH.


For comparison, a 100W rated "bookshelf" 8 ohm speaker , dimension 16 x 10 x 8" with 5.5" woofer dia.


BRAND TYPE W RATED RES 1 kHz R L

DALE RH 250 10 9.9 505 MH

Ohmite WW 225 10 10 55 MH

Ohmite REHO 150 10 10 180 MH

Ohmite 150 10 9.6 100 MH

SPEAKER TC 100 8 21 2.61 MH

I was surprised at the high L of the WW resistors and rheostats and the low L of the speaker.

We could post photos later if anyone is interested.

Bon journee,

Jon

hello part C765 , C766 .0022UF 250VAC? I found them on rscomponet?? https://it.rs-online.com/web/p/condensatori-a-film/0210572
the other two 1000uf 250vac
are electrolytic shouldn't be hard to find
--
Roberto

hello Dave thanks
i measured the battery yesterday and 3.45v ok for digikey but its better for me if i find it locally i saw it on ebay in germany they sell around 11 dollars new for replacement i saw put an additional 3v battery before taking it off which is the simplest method by viewing several videos on yotube For the capacitors tell me which model to buy?
?
--
Roberto

First of HP used to build instruments sometimes for decades, the 8660 for example was built for almost 20 years starting with the A and ending with the D model , the changes mostly were applied to the front panel.
As far as the testset concerns HP sometimes designed, build and sold accessories that were meant to support the service and/or calibration of a certain instrument, in this case the accessory was branded with the model number of the unit it was supposed to support followed by a letter and 2 digits, for example 8660C-K10 .
Altough this systematic can be a bit confusing it was quite common with HP back in the pre Agilent days!

I have just recently started collecting old Hewlett Packard test equipment, though I have been actively collecting and restoring Tektronix equipment for over a year. I started my HP collection with several of Nixie display counter-timers (5245L, 5246L, and a 5326B), a function generator (8116A) and a bunch of bench power supplies (6211A, 6212A, 6216A, 6217A and 6218A) and have been really happy with everything. My most recent acquisition is an 8660C synthesized signal generator, and while searching for a service manual I ran into a confusing situation that I have seen before with HP equipment: the model number appears to have been reused for at least one other instrument (in the case of the 8660C it is also the model of a phase modulation test set).

Unlike other model number collisions I have seen, these two instruments are not separated by decades: the 8660C test set is from 1975, while the 8660C signal generator is from 1981. Is there some explanation for why HP would reuse instrument model numbers like this?

-- Jeff Dutky

开云体育

Hello,

I did not ask this question earlier to be argumentative but only for my own curiosity.? When I was in high school (It was a very dark cave!), during the golden age of tube hi fi amplifiers, we used to routinely test them with an eight ohm 20 watt wire wound resistor as a dummy load and volumes turned down to avoid torturing other students.? As solid state amps were novel and unique then we paid special attention to always loading them similarly to avoid damage to output transistors.? I made a post on another forum to an obvious novice to make sure he also put on a similar dummy load and was rebuked for having suggested such a thing.? I let it drop but still question whether a wire wound load resistor is significantly inductive to impact audio frequencies.? I look forward to getting smarter.

Thanks,

Jack

Answer...u,clear the intended frequency range, so type in l8ad.is in question.

1/ no need to vary the load resistance as hi-fi amplifiers are rated driving a 4 or 8 ohm resistance load, not a voice coil.

2/ Depending upon the rehostat watts and total resistance and curve, the self induction varies considerably, I guess 5....100 uH, I can measure a few.
Speaker voice coils ,are smaller cross section, fewer turns, but in a strong magnet gap,so, very different induction.

3/ Many power amplifiers have the transformer secondary drive the speakers through a series LR, so situation is not as simple as one may like.

4/ One can easily find Dale non inductif wire wound power resistors.
Alternative is DIY non inductif load from a carbon rod and some band-it clamps.
However at low frequency eg 500 Hz the rehostat or induction resistance is OK

Bon chance

Jon

On Thu, Feb 24, 2022 at 05:02 PM, Froggie the Gremlin wrote:

1/ use a non inductif 8 ohm power resistance NOT rehostat for load. Too inductif!

I haven't done the math or measurement, but I question that (for example) a 20 ohm, 20 watt rheostat has enough inductance to be important at audio frequencies. It's an air-core coil, wound on ceramic, and only a few dozen turns or so.
And note that for many measurements the rheostat tap would be below half, so even less inductance.
A set of fixed R's is fine of course, but it would be really nice to have a continuously variable setting, so that (for example) you learn that the optimum power occurs at 6.7 ohms.
If I had a bunch of audio out xformers and an 8903B, I'd be very tempted to build the test setup myself - it would be fun and educational to try different transformers, tubes, and voltages. [If you build it, include extra sockets for the output tubes, so that you can plug in various tube pin-outs easily. Or maybe use a switching setup, similar to a tube tester...?]

Pete

On Thu, Feb 24, 2022 at 02:31 AM, Attilio wrote:

Jim Allyn, I only have these transformer data:
Push-pull transformer for EL34 valves.
Secondary impedances: 4 and 8 ohms.
Maximum power: 40W.

That's probably all the information you need to make approximate measurements of the transformer frequency response.? As others have said, if you want precision, you pretty much have to build the amplifier.

开云体育

Why would the inductance of either the rheostat or a wire wound load resistor be of significance when it is basically substituting for a voice coil (a coil of wire)?

Jack

Atillo,

1/ use a non inductif 8 ohm power resistance NOT rehostat for load. Too inductif!
Carborundum made "globar" tubular loads for dummy loads for TX.

2/ Use classic RCA valve app notes 1950s on calculations of equiv plate resistance for various valve power amplifier types eg class.A, AB, AB2. The starting points are the valve spec sheets, eg EL34, 6L6, 8060.

3/RE compare transformers for audio output...no need as optimization demandes the larger heavier, lower resistance P and S, and thinner laminations, Finally even better, costly tapewound cut core transformers.

4/ Expect $75..$500 cost for a true high-end transformer, not a junker from China.

Bon chance

Jon

开云体育

Hi Alex,

I wouldn't describe that output stage as a push-pull topology. The pulling up is just passive; there's only active pulldown.

The output stage is a slightly modified, but otherwise plain-vanilla diff pair. The modification is the stacking of output transistors to increase breakdown voltage. Unlike a cascode, the upper transistor of each stacked pair is biased by a fraction of the output voltage, rather than by a fixed voltage. This way, the two transistors see roughly equal VCE and the stack acts like a single transistor with twice the breakdown voltage.

Check everything along the signal path, including common-mode values. You'll see something goofy at some point. It should be straightforward to identify the problem then.

--Cheers
Tom

-- 
Prof. Thomas H. Lee
Allen Ctr., Rm. 205
350 Jane Stanford Way
Stanford University
Stanford, CA 94305-4070

That is certainly what it looks like, but looking at the schematic the vertical output amp is not like any push-pull driver I am used to seeing.

Thanks Jon,
you are right, but the measurement can be used to make comparisons and to have a first indication of the transformer.

Pete,
great idea, I had thought about it, but I had not thought about the rheostat on the secondary to adapt the transformer to the final tubes of the amplifier.

-- Cheers
??? Attilio

Just my 2 cents here, but to me it looks like crossover distortion, so
you might want to check around the base polarization of a final
push-pull amplifier or a push-pull driver. Of course could be a
malfunctioning integrated (operational) amplifier.
I have the 1727A too btw
Frank IZ8DWF

The input signal for the captures was not the calibrator, it was an external function generator. I chose triangle because it makes it easiest to see the distortion on the screen, sine shows it too and obviously square does not, hence the external function generator.

I completely forgot that I had my scope in AC coupling for those captures. new ones are attached along with pics of both sine and triangle on the screen, which I just realized I never attached to my original post, oops.

Attached in order of pins 3/6, 4/5, and 12/13 of U1(hopefully). I also captured the final vertical output(AC coupled so you can actually see something).

Reminds me of a saying...
It is better to don sandals than carpet the world.
What's next?? Demanding people wear certain things (that has nothing to do with covering indecent exposure) in public or face jail time and/or fines? Oh...wait...

The following files and folders have been deleted from the Files area of the [email protected] group.

• /Above 10000 to 89999/50000 to 59999/54720A/54720A_FW_EPROM_DUMP.rar

By: boddax@...