Moving-Magnet Head Amp - without RIAA

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Has anyone published a schematic for a Moving-magnet head amplifier that merely loads the cartridge correctly and boosts the level, but without applying any RIAA equalization?

I understand that this sort of thing is common for Moving-coil cartridges, where a step-up transformer is either too expensive or undesired for other reasons. Such a Moving-coil head amplifier is designed to be connected to the input of a MM stage with RIAA.

I listen to all music through an excellent DAC, but there is still some music in my collection that's only available on vinyl. I have written my own RIAA equalization plugin in the digital domain, and my DAC also implements optional RIAA equalization in its DSP, and I hope this explains why I want to avoid the analog RIAA circuitry.

I've gone through an evolution of several stages so far:

1) Since I needed an adapter cable from RCA to 1/4" TRS or XLR inputs on my computer audio interface, I started by building in the 47 kΩ load in the adaptor cable. I started with the specifications for the input impedance of my audio interface, and calculated the resistor value that I needed to add in parallel to end up with 47 kΩ. This setup worked well, but the ground lug for my tonearm kept coming loose, as evidenced by obvious hum until I wedged it in somewhere.

2) After the ground lug came detached too many times for my patience, I decided to buy a cheap phono preamp with an actual screw terminal on the case. I then gutted the original RIAA circuit, measured the mounting points for its PCB, and fabricated my own passive adaptor. This time, I also added a capacitor to provide the 200 pF to 600 pF load specified by my cartridge manufacturer. The PCB is available at OSHPark. This worked well, but when I moved the cable to a different audio interface and again when I upgraded the electronics in the audio interface, the passive components are no longer loading the cartridge correctly.

3) My goal now is to build an active phono preamp, but without any of the RIAA curve. This circuit would be tailored for my specific cartridge, and the output would be at standard -10 dBV unbalanced and +4 dBU balanced levels so that any input on my various audio interfaces would work (without being fragile due to parallel resistor and capacitor combinations). I'm trying to fit everything in the same phono preamp case that I have, since the grounding screw is going to be a necessary component.

4) In the future, I hope to upgrade from MM to MC, and might just build Douglas Self's Moving-Coil Head Amplifier. I assume this could be connected directly to my audio interface instead of a MM preamp, and I could then continue to realize RIAA EQ in DSP.

By the way, I recently obtained Self's "Small Signal Audio Design" book, and this probably has everything I need. However, there are a few questions I have about my slightly atypical application that aren't directly anticipated in his text.

Despite the fact that I've not seen anyone talk about this, it seems like modern listening with DSP available might benefit from such a flat-frequency-response phono preamp. On a related note, I'd much rather archive my rare vinyl via such a flat preamp setup and then apply the RIAA curve during playback, rather than having to invest thousands into an analog RIAA EQ that doesn't suffer from the usual noise and component tolerances.

Brian

p.s. I'm planning to design around a +34V single-ended supply, since I'm a bit too cramped for space to pull off a +/-17V bipolar supply. To keep things simple, I'm using an AC wall-wart, but those have only two conductors - no center tap. Obviously, this power supply requires a virtual ground and capacitive coupling. Douglas Self recommends always using capacitive coupling on the cartridge connection to protect the cartridge from DC current. That seems like sound advice. I'm not too worried about the AC coupling on the output that my unipolar supply requires, since I'll probably use a subsonic filter that would defeat any potential advantage of DC coupling on the output anyway.
 
Member
Joined 2014
Paid Member
Scott Wurcer published a few schematics for that and there is a kit Wayne Kirkwood designed. There is also a thread I can locate where phantom powered MM/MC is discussed if that is of interest. Focus specifically on interfacing into the sort of mic inputs that people have on home studio ADCs



You can take pretty much any phono stage and remove RIAA of course, adjusting gain to suit.



Many choices :)
 
There is also a thread I can locate where phantom powered MM/MC is discussed if that is of interest.
I got excited about phantom power when I started this project, but then realized that I have the audio interface model with line level inputs and therefore no phantom power!

I would still be interested in looking over that thread, since I plan to power from +34V and that should be quite close to powering from a nominal +48V.
 

PRR

Member
Joined 2003
Paid Member
....the output would be at standard -10 dBV unbalanced....

At what frequency??

If we define the RIAA-corrected response to be "flat", then a non-corrected response will be 20dB down at 50Hz and 20dB up at 20KHz.

i.e. the bass will be down in the mud.

It looks like the top octave will be STRONG, but in fact most music (and anything you can get from a needle) has around 20dB fall-off above 1KHz.

That's before we get to the wide range of recorded levels on LP, which has no real "limit" like digital, or FM, or even tape; only a trade-off between loudness and time per side.)

Taking a rough guess: if RIAA gain of 100 at 1KHz gives a good nominal line level, take the caps and bass R out of the RIAA and it will probably give a tolerable level. For two common topologies:
 

Attachments

  • no-RIAA.gif
    no-RIAA.gif
    33.1 KB · Views: 930
A flat amp is going to require substantially more GBW and slew rate than one with active RIAA. I would use at least two opamp stages at 20 dB each. Maybe start off with NE5534As (910/100 ohm) and continue with a good FET input dual of highish slew rate (AD8620 w/ 4k3/470 ohm? I know they'll only do +/-13 V, but in this application a +/-9..12 V supply would be plenty).

Software / DSP RIAA is doable (and has been done before, even commercially) but you do want a decent ADC, and it goes without saying that you do not want to be fighting a ground loop issue in your setup, as the amp-DAC connection is going to be ~20 dB more critical in this regard.

As a compromise, I might try implementing the second gain stage as a good approximation of an integrator across the entire audio bandwidth. (Maybe inverting, 2k2 + 6.8n poly || 1M?) This characteristic would be easily reversed in software, and any parts tolerances would only result in some channel imbalance (easily enough brought in line either in software or by making one channel's gain slightly variable with a trimpot, applying a signal split to both channels to adjust) or frequency response deviations in the subsonic range. It would make ADC high-frequency noise a bit more critical, of course, but in the days of 100+ dB ADCs I don't see that as a major problem.
 
There is no reason why you can't generate +/- supplies from your wall-wart. I believe Doug Self shows how in his book. It's pretty easy and works well as long as the current draw isn't too high as each 'half' is being charged with half-wave rectifiers.
Thanks.

I've had success with the voltage doubling technique that Doug mentions in Chapter 25, under "Dual supplies from a single winding," but that was for a different project that required a +/-7V supply. It actually worked "too" well, in that the 12 VAC wall wart produced more than 34V peak-to-peak at the regulator inputs, which exceeded the 25V input rating of the positive regulator. The negative regulator can handle 35V to 40V inputs, so that one worked fine.

I'm a little reluctant to use half-wave rectification, because for some reason I assume that it would produce more noise. There'd be more ripple, and I'm already cramped for space with regard to capacitors. Maybe I need a bigger case!

All things considered, I think that my biggest objection to bipolar versus a unipolar supply is the extra components. I need a second regulator, and two additional smoothing capacitors for the input and output of that regulator. I'm already cramped in the case, having only 22 mm height for the capacitors, and the 18 mm diameter space that's available limits me to 1300 µF.

What's funny is that Self compares unipolar supplies with the caveat that extra components are needed in the form of DC blocking capacitors in some of the places where gain resistors might normally be grounded. In my practical case, those additional capacitors are much smaller in physical size than the extra capacitors needed for a negative supply.

I still need to SPICE up the circuit so far in order to estimate how much supply current I will need. That might answer a few questions about component sizing.

I'll probably design a second version of this No-RIAA Phono Preamp using a bipolar supply, but only after I've secured an affordable case that has in integrated phono ground lug screw.
 
You may be interested in this Revox B291 turntable schematic on page 6-14. Revox called it an impedance converter but it's just an RIAA preamp without the RIAA.
STUDER REVOX B291 SM Service Manual download, schematics, eeprom, repair info for electronics experts
Thank you!

That is interesting, but if I found the right page it seems to be an all-transistor circuit. I may not be satisfied with the distortion performance of an all-transistor design, so I'll probably stick with the NJM5534DD plus LM4562N design that I'm working on now.

When I upgrade to Moving-coil, I'm thinking of building Self's hybrid transistor plus op-amp preamp in Chapter 12. He describes a flat (no RIAA) head amp for MC that's designed for feeding a standard MM phono preamp. That would be almost ideal for my Metric Halo Labs ADC, although I might have to increase the gain.
 
rsdio said:
....the output would be at standard -10 dBV unbalanced....
At what frequency??

If we define the RIAA-corrected response to be "flat", then a non-corrected response will be 20dB down at 50Hz and 20dB up at 20KHz.
Your question is heading in the right direction. At the very least it's related to questions that puzzled me for a while.

Thinking from a practical sense, though, -10 dBV is basically setting an overall level, regardless of frequency. All of the frequencies in the total signal have to be within the peak-to-peak bounds of a typical -10 dBV input. I don't happen to know what level is considered to be safe from clipping for unbalanced gear, since they all seem to report RMS levels.

If my hottest MM cart puts out 8 mVrms, then I assume I have room for almost 32 dB of gain to stay within -10 dBV levels.

In other words, I think that we can safely ignore the levels at specific frequencies. As long as the total signal can be digitized by the ADC without clipping, then the RIAA curve can be applied in DSP. My RIAA plugin has an automatically-calculated gain that guarantees that the output levels will not exceed 0 dBFS (so long as the input levels don't exceed 0 dBFS - it's a floating point DSP).
Taking a rough guess: if RIAA gain of 100 at 1KHz gives a good nominal line level, take the caps and bass R out of the RIAA and it will probably give a tolerable level
I started with one of Doug's phono circuits and tried to delete the EQ caps while retaining the basic "gain" of the op-amp. I ended up with something around 29 dB to 30 dB of gain. When I later went back and compared the 8 mV output spec from the cart with the 316 mV nominal level of -10 dBV, I also got +31.9 dB, which is about the same.
 
Last edited:
A flat amp is going to require substantially more GBW and slew rate than one with active RIAA. I would use at least two opamp stages at 20 dB each.
That's an excellent suggestion. I was already thinking of using two stages, one at +30 dB for phono to unbalanced levels (whose outputs I may never use) followed by an unbalanced to balanced stage with just under +12 dB. That's a total of +42 dB. Since the gain will be conveniently set by resistors, I should be able to try your approach of more evenly distributing that gain. The caveats are that my unbalanced outputs would only have a single stage (but, as I said, I'll probably never use those outputs) in either case, and if I change the ratio between the stages then my unbalanced output would be weaker (although the balanced outputs could remain basically the same).

Question: Since RIAA boosts bass, doesn't the RIAA output require even more slew rate? I can see how the gain bandwidth might be relaxed by cutting the highs, though.

Maybe start off with NE5534As (910/100 ohm) and continue with a good FET input dual of highish slew rate (AD8620 w/ 4k3/470 ohm? I know they'll only do +/-13 V, but in this application a +/-9..12 V supply would be plenty).
At the moment, I'm planning on NJM5534DD chips (split stereo) followed by two LM4562N (each handling +/- for a single channel). The LM4562N has standard dual op-amp pinouts, so I can try the AD8620. The caveat here is that both of my op-amp choices can handle a +34V supply, but I'd need to scale that back before installing the AD8620.

Software / DSP RIAA is doable (and has been done before, even commercially) but you do want a decent ADC, and it goes without saying that you do not want to be fighting a ground loop issue in your setup, as the amp-DAC connection is going to be ~20 dB more critical in this regard.
I'm happy with my RIAA implementation, but the ADC has basically the same.
My ADC is the Metric Halo Labs LIO-8, which far exceeds decent.

Your ground loop comments are spot on. The whole instigation of this project was grounding issues. I've probably solved the grounding issues with the passive PCB, but now I'm seeing ultrasonic noise pickup, and I assume that going active will solve that.

I'm not entirely convinced that I "need" balanced outputs, but since the LIO-8 has balanced inputs, I figure it's worth a couple of extra LM4562 stages to get maybe 3 dB better S/N.

As a compromise, I might try implementing the second gain stage as a good approximation of an integrator across the entire audio bandwidth. (Maybe inverting, 2k2 + 6.8n poly || 1M?) This characteristic would be easily reversed in software, and any parts tolerances would only result in some channel imbalance (easily enough brought in line either in software or by making one channel's gain slightly variable with a trimpot, applying a signal split to both channels to adjust) or frequency response deviations in the subsonic range. It would make ADC high-frequency noise a bit more critical, of course, but in the days of 100+ dB ADCs I don't see that as a major problem.
That's a very interesting idea. I'm concerned that it adds another variable that I'll just need to compensate for.
 
Member
Joined 2014
Paid Member
In terms of gain you have to deal with the gearchange of thinking in dBV to thinking in dBFS. I find the easiest way to look at it is to make the 0dB ref off a test record equal to -23 to -26dBFS. So for the LIO-8 set to +22dBu for 0dBFS you are roughly talking 0dB record=0dBu into the ADC. Or 40dB.

The handy thing about this is that most ADC designed for studio use have a green light that comes on at -23dBFS to indicate signal present. Handy kwik kal there :)

That LIO is a lot hotter than most stuff I am likely to use, so I can wuss out with a lot less gain. I only need 40dB gain for LOMC.

BTW this article Scott wrote for Linear Audio is well worth the €2.99.

For those who don't like rolling PCBs/vero there is this Phono Transfer System Construction Information - Pro Audio Design Forum Balanced in and out with minimal parts count and the option of a switchable RIAA if you want that.
 
Good points.

If I were targeting a +4 dBu mixing console, then I would not bother to boost the gain so much that I'd need to engage the attenuating pads. That sorta defeats the purpose because you have to add resistance.

The LIO-8 has digital controlled gain, at least as far as I understand, so it seems like almost any gain that falls within the range will be good enough. I'm quite familiar with setting the input gain levels for vinyl sources on the ULN-2 and LIO-8, and I always leave plenty of headroom to avoid any chance of clipping at 0 dBFS.

I'm more thinking of dialing in a nominal +4 dBu signal level rather than going for the extremes (high or low) of what the Metric Halo Labs MobileIO can handle.

Thankfully, gain is the one thing that should be relatively easy to change since I'm designing for through-hole metal film resistors. I mean, I also have soldering tweezers for thin film resistors, but I don't know that I'm cramped for space quite enough to incur the cost of SMD assembly.
 
BTW this article Scott wrote for Linear Audio is well worth the €2.99.

For those who don't like rolling PCBs/vero there is this Phono Transfer System Construction Information - Pro Audio Design Forum Balanced in and out with minimal parts count and the option of a switchable RIAA if you want that.
Thanks! Those threads look very promising due to the similar goals.

I got the impression that Scott's Linear Audio article is separate from the Pro Audio Design Forum, but I didn't see a link or full name. Did I miss something?
 
Member
Joined 2014
Paid Member
Yes, two seperate designs, both for similar goals. Scott's work is interesting for people who happen to have an ADC at home that has phantom mic inputs and opens up some interesting options for ripping vinyl using a battery powered field recorder. No ground loops guaranteed!


I even have one thing to try with a focusrite mic input where you just load down the cartridge to get the 75us pole passively before you reach the amplifier. This will increase noise but for fun worth trying.



worth checking at what gain your LIO has the lowest noise and target that.
 
Yes, two seperate designs, both for similar goals. Scott's work is interesting for people who happen to have an ADC at home that has phantom mic inputs and opens up some interesting options for ripping vinyl using a battery powered field recorder. No ground loops guaranteed!
I'm probably not trying hard enough (haven't searched "Linear Audio" for articles by "Scott") but could you at least mention his last name? An actual link would be awesome, although I gather that it's not a free download.

worth checking at what gain your LIO has the lowest noise and target that.
Excellent suggestion. I'll look into it.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.