Vintage Mic Transformer for MC step-up - gain-phase and impedance measurements

[trobbins]

In the early 1960's in Australia, Zephyr seemingly sold many of the type 130M mumetal cased microphone transformer. I appear to have four of these units, although they are not marked with a model number. I've started to measure their performance for MC step-up applications, as they can provide either +20dB, +24dB or +28dB gain. With the secondary loaded by 47k with 150pF shunt, the raw frequency response shows a +4dB high frequency peak out at 33kHz, which is flattened with an added shunt 39k+330pF RC to give a nominal -3dB bandwidth from 10Hz to 45kHz. The 24dB and 20dB connections have similar bandwidth, although the added RC requires tweaking for similar flat responses

The impedance spectrum seen from the primary winding is nominally 50 ohm, as anticipated by the advertised specs. Similarly the other connections present nominal 100 and 300 ohm midband.

The frequency spectrum was generated by a PicoScope 4224A using FRA4PicoScope software, with a 10mVpk sine driving primary directly from the scope signals output. The 150pF shunt loading comprised 57+17=74pF from the 1:1 probe and scope input, along with an added 68pF ceramic. The software can smooth out the noise, but only from 10Hz and above, so the plot is a little ragged, and there is no incentive to raise the drive signal level. The signal generator of the scope has a 600 ohm rated output resistance. I'm unsure how much influence the generator's output resistance would have on the character of the frequency spectrum - any thoughts?

The impedance spectrum was generated by REW software and an EMU 0404 USB soundcard and calibrated test jig. The added RC on the secondary winding influences the higher frequency character of that impedance plot, and contributes significantly to the dip in impedance in the 3k to 20k region. Any thoughts on how much of an influence the varying impedance would be, and whether it could be flattened along with also keeping a flat secondary frequency response ?

I'll play around a little more over the coming days.

Ciao, Tim

 

[moamps]

I believe that the best approach would be to simulate the real conditions in which that transformer will work. At least that's how I tested transformers for MC frontend. It should be assumed that the MC source has a low impedance (few ohms to several tens of ohms) and that the optimal load is in a wide range from 100 ohms to several tens of kilohms. In your case, I would connect, for example, a 10 ohm resistor to the primary of the transformer, and load the transformer on the secondary side with a 10k-47k rheostat, so that on the primary side you get a load for a 1:10 ratio of 100 to 470ohm.
If you set the generator to approx. 60 to 600mV before loading with the transformer, the signal at the input of the transformer will be around 1 to 10mV, which is more than enough voltage for measurement at secondary side and not too different compared to the actual situation when the transformer is loaded with an MC cartridge.

 

[trobbins]

Ta. Yeh my next set of testing was going to look at driving options.

I'd previously checked out what others had done, and perhaps the most informative view had come from Rothwell Audio Products who have an explanation web-page, but also some performance specs with their 3 products. That was why I settled on initially testing with a standardised 47kohm // 150pF load, given no other preference as yet.
mc step-up transformers explained
SUT

 

[Cqwet Dbdfte]

Use a test record, as your signal source has its own characteristics.
Load the seconadry with your preamp input, don't be afraid to change whatever loading it presents, yea, measure the impedance on the transformer primary connected to your preamp.
The cartridge mfg. should have some recommended numbers. Some like the cartridge more damped, some less, can change the character of the music a bit.

 

[Gold]

Some like the cartridge more damped, some less, can change the character of the music a bit.

With MC carts it has been my experience that this is a balancing act. Usually as soon as you get flat frequency response the cart starts sounding ‘constricted’. When you lighten the load you get a resonance peak. I usually tune them to have a slight resonance peak but sound more ‘alive’.

I also have the benefit of being able to compare input to output on the lathe.

 

[abbey road d enfer]

Any thoughts on how much of an influence the varying impedance would be, and whether it could be flattened along with also keeping a flat secondary frequency response ?

Increasing the load resistance also increases the reflected impedance below resonance, but decreases the impedance in the resonant zone.

 

[Gold]

Increasing the load resistance also increases the reflected impedance below resonance, but decreases the impedance in the resonant zone.

I may have the scenario backwards. I’m semi dyslexic and don’t do it that often. I’m going to be tweaking out an Ortofon SPU in the near future. I’ll pay attention to what I do.

 

[Gold]

I’m planning on using a Lundahl 1681 for the SPU. I’m using a 1678 for my Denon DL103 and am very happy with it. The 1678 has an amorphous core. I didn’t like the amorphous core on a Lundahl input transformer. The 1681 has a mu metal core.

Lundahl makes some large and expensive MC stepup transformers. Since the signal level is very low and there is no low end to speak of,because of RIAA encode, the added core size can’t much matter for distortion. The main issue I imagine is interwinding capacitance? Does the larger size help with that? Besides bigger and more impressive looking is there any technical advantage?

 

[Cqwet Dbdfte]

You can "look" at the transformers made by Ortofon and Denon to get a clue about needed core size.
Jensen also makes transformers suitable for this, in various step up ratios.
Magnetic shielding, not just electrostatic shielding is very important on these transfomers. The ubiquituos Altec 15095 have insufficient shielding, which I put on the junk pile, in favor of Westrex P-7884's which have double shields. The 15095's may work if placed inside a suitably magnetically shielded enclosure.
A 500uV dignal at 20Hz does not have much energy to saturate a core.

 

[Gold]

I guess my question is, does the increased size of the expensive Lundahl’s like the
https://www.lundahltransformers.com/wp-content/uploads/datasheets/1943.pdfhave any technical advantage over the regular https://www.lundahltransformers.com/wp-content/uploads/datasheets/1681.pdf

The data sheet mentioned better shielding. These will be in a shielded box so I’m not too concerned with that. The shielding on the small size 1678 I’m currently using is great. It’s dead quiet.

There are a few MC transformers I’d like to check out including Jensen and Tamura.

 

[abbey road d enfer]

Larger core increases Al so allows less turns for a desired inductance. Less turns means less capacitance, less resistance and less leakage. All these benefit HF response.

 

[Gold]

Larger core increases Al so allows less turns for a desired inductance. Less turns means less capacitance, less resistance and less leakage. All these benefit HF response.

Thanks Abbey.

 

[Cqwet Dbdfte]

...the data sheet would spell out the frequency/magnitude response and max power levels. Unless you are designing a transformer, the rest is academic.

 

[Gold]

...the data sheet would spell out the frequency/magnitude response and max power levels. Unless you are designing a transformer, the rest is academic.

It would, except it doesn’t.

Flat high frequency response is always the problem with phono cartridges. At least there is a technical basis for why the more expensive one might sound better. I’m not going to pay for silver wire but maybe the bigger core size is worth it.

 

[Cqwet Dbdfte]

All of these seven (7) MC transformers have datasheets, typ -1dB @ 60KHz. Looks pretty flat to me.
https://www.jensen-transformers.com/transformers/moving-coil/They recommend a 10:1 transformer, -1dB @ 100KHz, for Denon's DL-103 "20-45KHz" . The one setup I heard with DL-103 in in a Jadis setup was not stellar, kinda dull. Maybe not enough gain the in the Jadis phone stage? The Goldring MM E3 sounded great in the same.
After trying both 10:1 and 21.5:1 transformers I prefer the latter with my Ortofon HMC-10. (Not sure I would buy another MC after hearing the E3.)
In the quadro-phonic "CD-4" era some very high frequency vinyl was used, regular vinyl, not so much. The RIAA compensation would eat most of above 20KHz, like -20dB from 1KHz (if there was any) great if you have bat ears.
The Ortofon SPU +/-3dB 20-20KHz, Their recommended transformer ST-7 (16:1) 15-45KHz @-2.5dB, so not as high frequency as the Jensens @ 180KHz -2.5dB.

 

[Gold]

It’s the resonant peak in high frequency response of cartridges that is the problem. Not the bandwidth of the transformer. Limited bandwidth is a benefit along with superior CMRR.

I’m thinking the larger core may produce smoother high frequency response. Less peaky and resonant.

I use a DL-103 every day. I can compare the input to the lathe to the reproduce with the DL-103. A/B style. When properly set up it has very flat frequency response. There is some weirdness around 10k. Not more or less but not exactly right. The 1:10 transformer ratio is one of the nice things about the design. It basically sounds best with a 1:10 stepup terminated with 47K ohm.

The DL-103 has a conical stylus so has a different character at the inner diameter. Duller wouldn’t be wrong compared to an elliptical stylus but there are other things I like about conical styli.

I’ve tried quite a few cartridges at this point and tried to make them sound good too.

 

[Cqwet Dbdfte]

If you determined the resonant peak, with the cartridge as the cause, is this a loading effect?
Have you tried a custom loading for the output of the transformer and dug into the pre-amps loading circuit? The input circuit may have an inductor to tune out RF.

 

[Gold]

If you determined the resonant peak, with the cartridge as the cause, is this a loading effect?

Yes, you can hear more than the resonant peak change as loading changes.

Have you tried a custom loading for the output of the transformer and dug into the pre-amps loading circuit? The input circuit may have an inductor to tune out RF

Yes, I always try to tweak out phono cartridges. Now I use the KA Electronics Phono Transfer System. I have a custom preamp front end for each cartridge.

 

[MisterCMRR]

From an electronic point of view, phono cartridges are a very inexact science. First, you have the mechanical resonant peak in output due to the effective mass of the stylus plus coils or magnet working against the elasticity of the vinyl groove walls. Next, with high-impedance cartridges (moving magnet) you have the electrical resonance of the cartridge inductance and its total capacitive load (cable plus preamp input). Since both resonances (mechanical and electrical) tend to occur in the last octave (10 kHz to 20 kHz), the overall response depends on how the electrical resonance is "tuned" to complement the mechanical for flattest response in that range. For example (I think I recall the numbers correctly), the Shure V15 specified a load of 47 kΩ and 200 pF to get its published response.

Since moving coil types have extremely low inductance, the electrical resonance is way above the audio range and not a determining factor. But now the electrical resonance of the step-up transformer's secondary becomes the "tuning" point. The capacitive load on the transformer is affected by the cable (between transformer and preamp) and the input capacitance of the preamp. At one time Jensen made a boxed stereo MC transformer that had selectable input resistance and capacitance for that purpose. If you had a good test record (they have a very limited lifetime in terms of number of plays), it was possible to "tune" the system accurately rather than relying on listening tests.

To answer a previous question, core size and material have vanishingly small effect on high-frequency response. Like all transformers with a high-impedance winding, the distributed capacitance of the winding dominates the HF response. In Jensen designs, they are wound with what's called a "pilgrim step" winding technique to significantly reduce distributed capacitance. It's why the upper -3 dB point of the Jensen models are higher than most. Pilgrim step windings are commonly used on multiple pi RF chokes (as shown in the attached image) to raise their self-resonant frequencies.

 

[Cqwet Dbdfte]

You could "divide and conquer" by splitting up the issue, like driving the phono input from a signal generator, thru the transformer, or two of them back to back if generator does not have 5 ohm output ;-).
Once the preamp is squared away, you can work on the cart's loading to find a solution. Maybe a series R-C network on transformer sevondary.
I briefly looked at the schematic for that KA kit, a which shows some 49k resistance to ground, with a few choices for capacitive loading. This may not work out to cart's spec loading.
Like CMRR stated there are quite a few variables in mechanical damping and capacitive loading, in the cart but also tonearm, which could only be resolved with a test record.
I doubt there is a universal solution transformer that would work with standard 47k MM inputs.