Phono preamp headroom - why?

[TriodeLuvr]

I have now reworked the output stage in my current project and compiled the following measurements for maximum output voltage:

20Hz - 27V RMS (32mV input)
1kHz - 27V RMS (310mV input)
20kHz - 17V RMS (1.86V input

Maximum output voltage is the same at 20Hz and 1kHz because it is being limited by the capability of the output stage. Maximum output falls at 20kHz because the higher RIAA EQ attenuation causes the input stage to clip first.

Based on the Shure chart in post #7, the maximum input numbers above are all much higher than levels expected to be seen on a LP. This assumes a MM cartridge characterized by 5cm = 5mV with optimized loading.

Are there no reviews of tube-type preamps on ASR for comparison?

Jack

 

[atmasphere]

Nothing - it just makes sure that the headroom is maximised without negative consequences in case of clipping.

Absolutely yes. Still one should apply RIAA EQ in the first gain stage of MM pickups to optimize the headroom. What's wrong with this?

If the RIAA characteristic is applied prior to the first stage of gain you will have greater noise in the resulting circuit. With a MM cartridge this might have low consequence, but won't deal with the resonance, which should be eliminated in all cases for proper reproduction. Most of the studies of LP playback measurements I've seen pay no attention to this at all, clearly bad science IMO. Inductance in parallel with capacitance and its effects is electronics 101 in the first week...

 

[LTig]

If the RIAA characteristic is applied prior to the first stage of gain you will have greater noise in the resulting circuit.

I'm not talking about RIAA EQ prior to the first stage (who would do this anyway, that's a really stupid idea) but within the feedback loop of the first stage, like almost all application notes of opamp usage in phono preamps show as example.

The other way, to use passive RIAA EQ between to linear gain stages, lacks headroom and SNR, hence also not a wise design decision.

For MC phono stages a first linear gain stage makes more sense because it's easier to implement several gain settings due to the much wider range of pickup sensitivities. Headroom is not much of a problem since the input level is much lower and MC pickups don't suffer from high peaks in the first place. I never had any problems with my DIY MC phono preamp, sporting a 1st linear 20 dB gain stage, a passive lowpass (75 us), followed by a 2nd active gain stage handling the remaining RIAA EQ. It was dead quiet and ticks and pops were much less audible compared to the phono stage of my old preamp.

With a MM cartridge this might have low consequence, but won't deal with the resonance, which should be eliminated in all cases for proper reproduction. Most of the studies of LP playback measurements I've seen pay no attention to this at all, clearly bad science IMO. Inductance in parallel with capacitance and its effects is electronics 101 in the first week...

Yep. Proper loading fixes this. At least in the past makers of MM pickups specified the proper load (resistance - usually 47 kOhm - and capacitance) for optimum frequency range in the highs, which also flattens the peak. I don't know whether such specs are always given today - I would not buy a pickup without a full set of specs required for proper loading and arm compliance.

 

[atmasphere]

At least in the past makers of MM pickups specified the proper load (resistance - usually 47 kOhm - and capacitance) for optimum frequency range in the highs, which also flattens the peak.

Apparently not as much as it should! So often the actual correct load is something quite different.

When you say

within the feedback loop of the first stage, like almost all application notes of opamp usage in phono preamps show as example.

-do you mean if you have an input transistor or tube, that you would use the RIAA EQ wrapped around that (like an opamp), rather than two stages of gain?

 

[TriodeLuvr]

At least in the past makers of MM pickups specified the proper load (resistance - usually 47 kOhm - and capacitance) for optimum frequency range in the highs, which also flattens the peak.

Discussion at the link below appears to indicate that manufacturers have often recommended values on the basis of convenience rather than optimal response. Seems to me that the only way to know is to check resonance and high frequency response with a test record.

Cartridge Loading Explained

Jack

 

[LTig]

-do you mean if you have an input transistor or tube, that you would use the RIAA EQ wrapped around that (like an opamp), rather than two stages of gain?

With a single transistor or tube there is probably not enough loop gain for a full RIAA EQ within the feed back loop. Traditional circuits using 2 transistors do full RIAA EQ though.

For MM preamps there is no need to use a transistor circuit when 5532 opamps are available dirt cheap with low enough noise and distortion. MC preamps can be done with 2 opamps but at a cost (AD797 in stage 1) - fine for DIY (like mine) but not so much for high production numbers.

 

[TriodeLuvr]

I think there's a significant problem with the testing being done here on ASR regarding preamp headroom. It's not the test procedure itself, which appears to be scientifically performed with good test equipment. The problem is that applying these results as a measure of a preamp's quality, particularly when accompanied by comments aligning it with the intensification of pops and clicks, is simply without merit. In order to make such a judgement, one would need to know much more then the maximum tolerable input voltage. The maximum output level is also critical, because a preamp with a comparatively low maximum might fair more poorly in the test, yet still be capable of greatly surpassing the permissible input level of the amplifier that follows it. Another very important aspect of this has to do with exactly where in the circuit the preamp is clipping or distorting, as well as the recovery time when that stage misbehaves. Unless the testing reveals headroom that is extraordinarily poor, so poor that the preamp cannot function for its intended purpose relative to recorded levels, the results are meaningless as a figure of merit. I have to wonder how many potential purchasers of a high quality product offering good value have been put off by these measurements and the critical comments attached to them.

Jack

 

[MaxwellsEq]

I think there's a significant problem with the testing being done here on ASR regarding preamp headroom. It's not the test procedure itself, which appears to be scientifically performed with good test equipment. The problem is that applying these results as a measure of a preamp's quality, particularly when accompanied by comments aligning it with the intensification of pops and clicks, is simply without merit. In order to make such a judgement, one would need to know much more then the maximum tolerable input voltage. The maximum output level is also critical, because a preamp with a comparatively low maximum might fair more poorly in the test, yet still be capable of greatly surpassing the permissible input level of the amplifier that follows it. Another very important aspect of this has to do with exactly where in the circuit the preamp is clipping or distorting, as well as the recovery time when that stage misbehaves. Unless the testing reveals headroom that is extraordinarily poor, so poor that the preamp cannot function for its intended purpose relative to recorded levels, the results are meaningless as a figure of merit. I have to wonder how many potential purchasers of a high quality product offering good value have been put off by these measurements and the critical comments attached to them.

Jack

As is always the case, simple summaries and conclusions miss out on nuance. You are correct in that no part of the preamplifier should clip when exposed to pops, clicks and very hot cuts. However, of course, there's no point having loads of headroom in the 2nd or 3rd stage if the 1st overloads, is unstable or takes a long time to recover from perturbation.

Most reviews look at the headroom of the phono preamplifier as if it's a black box. Therefore when stating a DUT has insufficient headroom, the review is not normally identifying the front end or any other stage, just the end-to-end headroom (e.g. the point at which a certain input increases distortion of the whole device).

 

[TriodeLuvr]

Therefore when stating a DUT has insufficient headroom, the review is not normally identifying the front end or any other stage, just the end-to-end headroom (e.g. the point at which a certain input increases distortion of the whole device).

I understand, but that number can't be used as a basis for qualifying the preamp's propensity for pops and clicks. That depends on exactly where it's clipping/distorting and the recovery time. This is not a small problem in a review claiming scientific merit, particularly as it will likely be used by readers in making purchasing decisions.

The second issue here is lack of notation in many of these reviews regarding the test frequency. It's meaningless to state that a phono preamp distorts at a specific input voltage without including this information.

Jack

 

[atmasphere]

I understand, but that number can't be used as a basis for qualifying the preamp's propensity for pops and clicks. That depends on exactly where it's clipping/distorting and the recovery time. This is not a small problem in a review claiming scientific merit, particularly as it will likely be used by readers in making purchasing decisions.

Jack

That's not exactly true. The overload characteristic has a big effect on whether it will generate ticks and pops.

 

[TriodeLuvr]

The overload characteristic has a big effect on whether it will generate ticks and pops.

Of course it does. But the overload characteristic is not what is being reported in these tests. That can't be conveyed with a single number. You know this, because you're involved with ampllifier design. You know that an amplifier that's direct coupled doesn't produce the same sound with a transient overload as one with RC coupling, even if they both clip at the same level. Presenting a single number as a figure of merit in this regard is meaningless and misleading.

Jack

 

[sergeauckland]

Of course it does. But the overload characteristic is not what is being reported in these tests. That can't be conveyed with a single number. You know this, because you're involved with ampllifier design. You know that an amplifier that's direct coupled doesn't produce the same sound with a transient overload as one with RC coupling, even if they both clip at the same level. Presenting a single number as a figure of merit in this regard is meaningless and misleading.

Jack

On the other hand, if the overload margin is sufficient, then any transients will never overload the amplifier, so its characteristics won't matter as they're never exercised. That's why a healthy margin is required.

S.

 

[atmasphere]

On the other hand, if the overload margin is sufficient, then any transients will never overload the amplifier, so its characteristics won't matter as they're never exercised. That's why a healthy margin is required.

S.

To be clear, the overload margin must exist at high frequencies in particular.

 

[TriodeLuvr]

if the overload margin is sufficient, then any transients will never overload the amplifier, so its characteristics won't matter as they're never exercised.

If only life were that simple. Please let me know when you find a preamp meeting this criteria. The overall structure of 60dB gain at 20Hz, 60dB EQ attenuation at 20kHz and a practical output limit of perhaps 2V or 3V at the input to the audio system pretty much guarantees that the goal you've described can never be met. Some transients WILL overload the preamp in one stage or the other. Again, describing only the input level at which this occurs is meaningless in terms of the audible consequences for the reasons I've already stated. It is exceedingly misleading to connect this test to a supposed propensity for accentuation of pops and clicks. I don't know how to say this more clearly.

Jack

 

[LTig]

If only life were that simple. Please let me know when you find a preamp meeting this criteria. The overall structure of 60dB gain at 20Hz, 60dB EQ attenuation at 20kHz and a practical output limit of perhaps 2V or 3V at the input to the audio system pretty much guarantees that the goal you've described can never be met.

This is not correct. A passive RIAA EQ attenuates 40 dB at 20 kHz and 20 dB at 1 kHz.

Some transients WILL overload the preamp in one stage or the other. Again, describing only the input level at which this occurs is meaningless in terms of the audible consequences for the reasons I've already stated. It is exceedingly misleading to connect this test to a supposed propensity for accentuation of pops and clicks. I don't know how to say this more clearly.

Lets step back and see what a MM phono preamp does as a black box: It amplifies 40 dB at 1 kHz, which means 60 dB at 20 Hz and 20 dB at 20 kHz. To get an output amplitude of 2V (comparable to digital sources) at all frequencies it needs an input voltage of 2mV@20Hz / 20mV@1kHz / 200 mV@20kHz. Anything higher than these levels and you risk clipping the following audio chain.

Using a standard one opamp (like NE5532) phono preamp with EQ implemented inside the feedback loop (active RIAA EQ) you can be sure that the input levels shown above do not clip the opamp. The NE5532 can probably output 8V without clipping so even with 8mV@20Hz / 80 mV@1kHz / 800mV@20kHz it will not clip, but surely the following chain must clip somewhere (eventually the power amp).

With passive EQ the situation becomes worse. A 60 dB linear gain 1st stage using a NE5532 (or similar with +/-15V supply voltage) will clip at 80mV at all frequencies, so this is off the table. A 40 dB linear gain 1st stage will clip at 800 mV, so is fine from this perspective, but then the passive RIAA EQ attenuates 20kHz signals back down to their input level, so the 2nd stage with a linear gain of 20 dB may end up with a noise penalty of up to 20dB compared to the active RIAA EQ shown above. You don't want that either.

 

[TriodeLuvr]

This is not correct. A passive RIAA EQ attenuates 40 dB at 20 kHz and 20 dB at 1 kHz.

Thanks, it is indeed about 40dB attenuation at 20kHz. That doesn't affect the other numbers I mentioned or points I've been making about these overload tests.

Jack

 

[sergeauckland]

If only life were that simple. Please let me know when you find a preamp meeting this criteria. The overall structure of 60dB gain at 20Hz, 60dB EQ attenuation at 20kHz and a practical output limit of perhaps 2V or 3V at the input to the audio system pretty much guarantees that the goal you've described can never be met. Some transients WILL overload the preamp in one stage or the other. Again, describing only the input level at which this occurs is meaningless in terms of the audible consequences for the reasons I've already stated. It is exceedingly misleading to connect this test to a supposed propensity for accentuation of pops and clicks. I don't know how to say this more clearly.

Jack

Rather depends on the size of the pop or click. Of course a large one will overload any phono stage, but those are virtually unplayable, so of little concern.

As to your output level of 2 or 3v, I don't know where you get those numbers from, but any half-decent phono stage will output +20dBu, many +24dBu, so some 12dB higher, so should provide at least 20dB overload capability at 1kHz, and some 40dB at 20kHz. I disregard phono stages with passive EQ, as that's a very poor way of doing a phono stage, and just asking for trouble.

S.

 

[sq225917]

AOTBE, a phono stage with more headroom is less likely to clip than one with lower headroom, therefore less likely to accentuate pops n clicks. Thats all. It says nothing about the quality of the unit, its just a desired characteristic.

Simon, 26db of headroom.

 

[atmasphere]

I disregard phono stages with passive EQ, as that's a very poor way of doing a phono stage, and just asking for trouble.

I've had excellent luck doing that with no overload issues at all. But I am using tubes which run much higher B+ Voltages than the typical Vcc+

It is exceptionally free of ticks and pops as it simply does not generate them. I'm very used to hearing entire album sides with not ticks or pops at all.

 

[LTig]

I've had excellent luck doing that with no overload issues at all. But I am using tubes which run much higher B+ Voltages than the typical Vcc+

Yep, with sich high supply voltages clipping is difficult.

It is exceptionally free of ticks and pops as it simply does not generate them. I'm very used to hearing entire album sides with not ticks or pops at all.

Ticks and pops results from used and dirty records. They are only amplified by the phono preamp, not generated. Badly designed phono stages may make them more audible by making those spikes longer due to extended recovery time after clipping.