Output Transformer Makes Sound "Pump"

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

thermionic

Well-known member
Joined
Jun 3, 2004
Messages
1,671
Hi,

I’ve fitted some Sowter 3603 transformers to the outputs of a 6-channel Phono preamp.

The preamp works fine and – although I haven’t measured the output impedance – I know it’s happy to drive loads down to around 200 Ohms without major issues (output drivers are BC461).

With an incorrectly terminated transformer, you usually expect to see “ringing”, creating spikes in the upper Khz region – famously noted by the Rupert Neve / Geoff Emerick story, where Emerick noted the artefacts from an incorrectly terminated transformer ringing at 54K.

In my case, I have the opposite… Instead of a rising HF response, an FFT suggests nearly 2dB roll-off by 20K compared to without the transformer. I have spoken with Brian Sowter and he thought the issue somewhat odd. He told me I shouldn’t need to terminate the 3603, but I could experiment with a 10K variable resistor if I wanted to.

When I listen to the preamp, the bass seems more bloated than without transformer (not that unusual – transformers in general are known for exaggerating bass), the HF slightly duller – but the most noticeable behaviour – from a subjective POV – is a kind of “pumping”, slightly like a poorly aligned Dolby NR system, but subtler…

I could be imagining this, but the HF seems to “pump” in sympathy with the slightly “bloated” bass… I can well imagine some people to like this sound for harsh-sounding pressings as it’s quite subtle, but I’d like to figure out what’s going on here.

The PSU is overrated btw, and could supply many preamps of this type.

My main issue is that the sound with the transformers seems to lack the “speed” you get when you bypass the transformers… Anyone got an idea as to what’s going on?

It should be noted that the preamp’s output is fixed at 300mV (from a 4mV cartridge), so there’s no chance of overloading the transformer - and it should also be noted that there’s absolutely no chance of any of the transformers being subject to DC (fed from a cap – DC measured) – all 6 preamps sound identical.

Here’s the Sowter data sheet: http://www.sowter.co.uk/pdf/3603.pdf


Thanks in advance.

Justin
 
This is just a WAG but do you have any problems with excessive LF content from warped or improperly centered records. The LF content is usually ignored by most systems but transformers can saturate at very LF which might modulate other signals riding on top of the LF envelope.

The cap coupled transformer should generate a HPF pole that should be helpful in removing LF depending on where that pole is set.

I don't have anything to offer regarding transformer speed. If you're seeing a roll off already at 20kHz it may be audibly impacting your signal path with that roll off and associated phase shift.

I'll second Paul's question regarding what load impedance are they driving.

JR
 
Hi Justin,

note that coupling cap and primary inductance make resonant circuit.
If driver can push 200 ohm load, it has to have rather small output impedance
So Lpri-C circuit is undamped, thus you could have (with certain values)
peaking at low corner frequency. If this xformer is ment for isolation
of HiFi equipmet, it could mean it is not designed for low driver impedance.

If you have a peak at lets say 40 Hz, driver will see quite low impedance around
that frequency. Thus drivers` output transistors would try to pump more
current (thus they strain more) if there is LF content in material. Also for
core saturation it is more important how much current are you pumping
trough pri.

OTOH, all this seems qite far fetched as explenation for behaviour
you are experiencing (that is, I wouldn put money on my explanation)

cheerz
urosh
 
Thanks to all of you for your thoughts :guinness:

Since posting the query, I’ve tried swapping the 3603 for an 8403 Sowter transformer. The 8403 sounds perfectly smooth, with no discernable “pumping” (I have yet to run it through the FFT, but it certainly sounds fluid – as opposed to “robotic”, i.e. “pumping”). I also tried some non-descript German transformers that came from a broadcast console: a touch more colouration than the Sowter 8403 – but definitely no pumping artefacts!

The load impedance would be 20K in the case of my soundcard for the FFT and a minimum of 10K for my amp (I can’t find the impedance on any of the ‘spec sheets… I haven’t seen a modern hi-fi amp with below 10K, although I suspect this amp is probably higher, anywhere up to 50K)

It’s certainly interesting that I don’t get the problem with the other 2 transformers, yet with the 3603, the issue is consistent across all 6 outputs… It should also be noted that the 3603 is recommended for lines below +17dBu, whereas the 8403 can be used up to around +27dBu… 300mV is nowhere near +17dBu…Hmmm….


Justin
 
If this xformer is ment for isolation
of HiFi equipmet, it could mean it is not designed for low driver impedance.

The Sowter 'spec sheet quotes 0.05% distortion at +11dBu with a 40 Ohm source impedance... Am I missing anything here?


J
 
Just to back-up what Urosh said above, the coupling cap and transformer primary form a series-resonant circuit, which gives a low-frequency peak. Below this peak, the response rolls off rapidly.

At the resonant frequency, the impedance seen by the driving circuit is very low, and current flow increases substantially.

Damping this resonance, or using a bigger coupling cap to move the resonant frequency further down, well out of the audio band, could be the answer.
 
Infrasonic filtering of any phono signal is also vital. A 3rd order (18dB/octave) Bessel filter at 16 Hz will do a good job. In practice that's near-equivalent to a 12Hz Butterworth filter section plus a 12Hz passive filter somewhere in the chain. You could also do 16Hz 3rd order Butterworth, but I like Bessel better.

Peace,
Paul
 
I too think that it is probably a resonsnce formed by the coupling C and the L of transformer primary.

also, how loud were you listening? acoustic feedback could be contributing as well, especially if no subsonic filter was being used.
 
Thanks again to all :guinness:

I substituted a 470u cap in place of the original 220u cap on 2 of the channels and compared them to the others last night. To be honest, it didn't make a great deal of difference, but I wonder if I didn't play the right record... The audibility of the resonance could be affected by source material.

I didn't get a chance to try an FFT - once I have, I will report back.

Cheers!

Justin
 
You have received advice to a) increase LF resonse, b) decrease LF respone, c) look for resonance at some low frequency.

You tried plan "a" and it didn't help, perhaps consider plan "b".

JR
 

Latest posts

Back
Top