78RPM Heaven - a discrete multicurve phono preamp for your shellac!

Status
Not open for further replies.
Jump to Latest
#21
Hmm... I don't think you can get variable shelving with your topology which is a bit of a drawback. I'd prefer to use a rotary switch personally. I don't think RIAA accuracy is too much of an issue for 78s TBH, as long as you are within 1dB of the curve then that's fine for this application. Don't forget the frequency response of the cutter heads and recording equipment wasn't exactly great back then.

I prefer rotary switches, really. And it seems your topology boosts subsonic frequencies with a 2.2uF cap? I can't imagine that would be good for the rumble rejection factor.

There's no right or wrong way really so if you want to build yours differently then that's cool! I personally prefer my approach. I don't understand this 'insertion loss' from R19 and R20, they make up the necessary input resistance needed for the classic inverting 'see saw' shunt feedback amplifier.

More of my discs use the NAB curve than RIAA so I think it's best to make sure it can replicate all the curves well enough while being quick to adjust and sound acceptable.
 
#25
I have been trying to come up with a very simple preamp using as few parts as possible. An EF 86 pentode for gain and a 3 transistor omp amp for 3180 /318 uS. The 75 uS by loading the Denon DL 110 with a capacitor. The idea is to have state of the art performance with as few stages as possible. The really good news is having 78 EQ would not be difficult. BTW the input impedance of the op amp is > 1 Meg. The speaker phase reversed to suit. The EF 86 I have are high grade Mullard. The CV types wire ended. It was mostly so before I no longer have need they are used. If like Leak that will be 65 V at the anode. Transitors BC327/337 and a JFET CCS called a CRD. The EF 86 is almost a zero distortion device and even has good gain if triode, the pentode also is supurb. Hiss is OK for my needs. The op amp gain about 16 at 1 kHz. It looks like a JLH power amp. Output current suits a > 10 K input. The EF86 could have a current source, MPSA92 would be OK. My OB speakers need a bit of EQ. It is tempting to do all in one EQ in the phono stage. The line sources then EQ'ed to suit. The preamp gain might need to be 500. even in triode I am about there. In pentode I can easilly do it. A Denon with a 78 stylus might be a good idea. It is 8 cu compliance so should be OK. Shure M44-78 also.
 
#27
It's still in my head. I will start with a simple NE5532 front end onto an ESP Audio idea. I will try to simplify that further. Example No 6. It states 2M8 input impedance. If I am careful the EF86 can drive that ( + 1M = 0M7 ). In triode it will. A gain of 100 might be enough. This begs the question if an ECC83 with current source could be used. The idea below is concept. The 1M to the op amp input might work. It can be set lower I am sure as 10 V rms as the worked examples is not required. The EF86 cathode cap comes later as will grid stoppers if required. The distortion of the set up will be about 0.1 % and a tad worse at 50 Hz. DC offset also, with + / - 20 V I doubt it matters. A pot could be added to the feedback arm to set it. As I will be looking at the amplifier input ( Quad 33 ) I might replace it's input cap and have none at the output of the preamp. I have doubled the gain of the Quad 33 and get very OK results. This is a speculation. The Quad for all of it's simplicity works well. I hated it when young because I never got it's gain adjusted to suit. On the Quad two resistors 470R and 220R changed to 510R and 180R suit the DL 110 very well ( circe 1mV ). If not the volume has to be on 10 sometimes. Looking at the circuit Radio 2 suits. This has a input level of 100 mV. A gain of 100 should suit it. A gain of 150 maybe better.

Opamp Alternatives

le25lB7.jpg
 
Last edited:
#28
Seems a little strange. If you are going to operate the EF86 in triode mode and you are building a 2 channel preamp then why not just use an ECC83?

If you are playing 78s I've found a variable low pass filter is indispensable. It's a good idea to have 5, 7, and 10KHz cut-off frequencies.

Let me know how you get on 🙂 .
 
#32
monty78pig said:
Seems a little strange. If you are going to operate the EF86 in triode mode and you are building a 2 channel preamp then why not just use an ECC83?

If you are playing 78s I've found a variable low pass filter is indispensable. It's a good idea to have 5, 7, and 10KHz cut-off frequencies.

Let me know how you get on 🙂 .
Click to expand...

I made a little negative progress with it. As I should have guessed a JFET op amp is a better next stage. Looking at condenser microphones they need the same so that's where I should go next. I have some TL072 and OPA2604 so can try ideas.

The Quad 33 has very OK tone controls which will serve well. The Quad is many times better than it should be. As I understand electronics better I can see it is not a stupid product. Perhaps it's passion was to save money ? I don't think so. It seems minimalist with excellent distortion. The Radio input has a rather high 100 mV sensetivity. The tape input pots it down. It sounds very good. I mistook the tape input for being a better and later stage. What I think I hear is finding an exact window of gain. The Quad is something like the old VW Beetle, nothing is wonderful except the ability to outperform the parts it has. By stretching that thought one arrives at the Porsche. I drove a VW-Porsche hybrid with a soft tune engine. It was sublime. The Quad is like the VW in that it only just works, it isn't hard to kill that it does. When you do it sounds dark, dull and boring. Or thin when too low.

The EF 86 idea might go even further than I intended. As a triode it seems a waste of a device except it is a very wonderful triode somewhere between ECC83 and ECC 82. When the gain is made pentode MC should be possible. This might mean 50 dB noise. I could live with that just to have something useful on hand.

Overall I need a gain of 150 ( I always think double the stated seems about right 150 x 1.6 = 240 mV ). That might mean at gain of 60 from the op amp reducing to 6 at 1kHz. The ESP op amp favours low impedance. I feel that the 3180/318 would need large values. Any off the shelf op amp could form a buffer. That throws away the simplicity except in the build. The MC option would offer very high gain in pentode. Even an Ortofon SPU should be OK.

One thing I think I have discovered. Most EF 86 preamps sound like the performers are standing in a room. Depending on the design this is highly coloured. Some EF86 are nothing like a Mullard. Looking at the specs the EF86 is a very good amplifiying device. A typical shunt feedback circuit should be the colouration heard, running out of HF gain and parts used make it worse. The next stage loading was not ideal in typical circuits. To my mind even 1 M ohm is too low if pentode. Not sure of the next part. Surly loading an EF 86 with a CCS makes the need for a high impedance next stage even greater ? Doubless it copes. Better it isn't asked to ? The VAS of a transistor amp is this way and it does cope ? A wackey CCS would be a ECC82 cathode follower and bootstrap to EF 86 anode load ( 2 x 150 K ? ). My instinct is that it is rather good as it is half way between passive and active. ECC82 is cheap so why not ? Most sound great. If using OP2604 I would set common grid/gate leak at 2M and hope for the best. I will draw this if you like? Seems easy to see without that. The old Leak preamps were feed via a very scruffy piece of co-ax the the power amp. It was very short to keep the capacitance down. Doubtless this gets replaced by some 300 pF per metre costs the Earth cable of a useful 1.5 metres? The sad fact is the lightly loaded EF 86 was the magic of the design. Replaced with something awful and dark. No doubt this is why cathode follower outputs were used? I don't like them much except in the bootstrap idea perhaps ? Of practical 78 preamps the Leak was the best I heard. The Quad 2/22 was not it's equal to my ears. A Darlington buffer I built was OK.
 
#33
For a reference point I built a modern all in one active RIAA stage today. Same thing as the Quad in a way except using NE5532 or MC33078. The Quad is better even though I had to increase the Quad gain 6 dB for it to be loud enough ( DL110 ). One thing I did that was clearly different to the op amp design is a 4.3 uS output filter on the op amp output. Very easy to do and stops the tizz, very odd if thinking where 4.3 uS starts to work. One thing I might do to the Quad is exchange it's resistor in the dics stage for CCS. That might be all I am looking for utill I get a better plan. This op amp preamp is 43 dB gain which might be a tad high. 40 dB should be enough ( 160 mV into 100 mv senstivity if so ). I am using 1nF and 3 nF polyester 63V ( 2 x 1.5 ). I think the polystyrene I normally use is part of why this one is only average sounding, I was hoping the difference would be less. I used 100 % DC feedback as the DC offset was high. Even so there is 222 mV on each output, strange it should be so equal . The input current is 10 mV into 47 K. Knowing how OK a Rotel RA 931 sounds with similar it has been fun to try. How the Quad does it with two transistors beats me ?
 
#34
RJM Audio - The Very Simple Phono Stage

The parts shown are what I had and not some great plan. The input stage of the Quad might alter the 4.4 uS a bit. The 220 uF is a bit large. Again what I had. 10 uF a bit small in polyester. Non polar caps are fine if below 0.4V as here. I must say not at all bad. As the Quad has polyester inputs ( 300 n F ) I didn't use an output cap. Hiss is very low. I used exact values by hand picking.

4PEBpAV.jpg
 
#35
Just in case someone should want to try this idea I have just done some small tweaks. R2 1K giving circa 36 dB. It's cap now 10 uF 100 V polyester ( 16 Hz - 3dB ). Output filter now 220R + 10 nF mylar ( 2.2 uS ). The preamp does address this problem. I feel the extra passive filter worth having. The Quad is quieter and more open. However it is fussy and needed some work to sweet spot it. The sound is like the budget amps, but a much better version. A bit closed in and shows S sounds. Glad I made it as predudice said don't bother. It was very easy to make and worked first time with no on scope problems. Will try putting the output of the op amp into class A a bit. It might be that alone that makes the Quad sound better. I think I have some 2 mA JFET's.
 
#37
I am doing a step by step test from least expensive to something like in your design hopefully with EF 86 input. The Class A trick using CCS was interesting on the MC33078 one op amp design ( gain of 36 dB using 1K gain resistor ). Sounding more like the Quad but not right, good stereo but no depth and all upfront. I reduced the current and form by going from 2 mA CCS to 22K pull down reistor to - ve rail ( circa 0.5 mA ). Much the same. I will try the cartridge impedance loaded to get 75 uS next. If this is a practical version 78 EQ can be looked at. MC33078 has 0.0022 % THD @ 1V output. That would seem to say still 0.1% at a gain of about 70 and 250 mV, not bad. This holds up to 50 kHz quite well and will be less due to no real output. The Quad seems to have it beaten on all counts. I must throw the text books away. What I guess they did was fine tune it. Some will say op amps are better on all counts. The hiss advantage of a BC109 type device over NE5532 type op amps is very obvious. The Quad is two transistors that are forced to have a gain of about 60 and do RIAA ( 600 at 20 Hz ). Every text book would say on every count do not think of doing it this way. With the resitors shown in red it even has enough gain for a DL110. The 33K collector load of the output BC109 could be a CCS. It is very odd it works so well. The gain change is enough to suit all LP's. The 510 and 180 R took many listening tests to get right. It is that critical. Too low a gain sounds dry and too much bloated. We are talking 6 dB to hear that. The overload margin is poor so I guess that's what we hear ? However it seems enough if careful. The Denon has output at 45 kHz so the Quad is being tested to the limit. The actual sound is rather good. Even the bass filtering is not too obvious. I suspect PU resonance at LF boosts the bass enough that Quad ESL would not like it ?

84iTlqo.jpg
 
#39
I made very good progress with the very simple design. 3180/318 as it was ( 316 K + 30 k + 10 nF polystyrene , 470 R gain, that is 36 dB @ 1kHz ). No 75 uS in the active part. I have shunted the DL110 with 470 nF to give that ( 160R + 380 uH ). This really works and is a transformation. It resembles a 1962 Tobey and Dinsdale idea. It would be very tempthing to convert the Quad circuit to the T&D. I like the Quad bootstrapping to be like an op amp, that helps . The lack of depth is gone in the now NE5532. The treble is better, it was OK before with 2 uS filter using MC33078. Now not required. For 78's the treble would be easy now. I have a spare DL110 for retipping.
 
#40
If this helps less ambitious builders this might be useful. Douglas Self improved the Lenco slightly as shown ( 24 V and symetry ) . It will give low distortion well above 20 kHz up to 1 V with a good headroom. I might be lucky with the DL110 that simple loading like this works OK. The current seems available to cope. The sound is better than active EQ. The 3180/318 is concept. It is to make it as near to the Lenco as possible so as to keep DC points. The open loop bandwidth is enough for 3180/318 uS , The 2.2 uS output filter is optional ( 220R 10 nF ). The input reisistor 3K9 is from the Lenco. I have built similar with MC33078 and NE5532. Both sound good, the MC33078 softer and the 5532 slightly glassy in a nice way. I have some OP2604 that seem to be hiding, shame.

The Quad 33 is more natural. My hunch is the Lenco also. Regardless of what Mr Self says 0. 1 % THD LF reducing to 0.01 % way past 20kHz is not bad if below 1 V rms . As I need 200 mV I should have almost none.

I have just bought a stylus for a JVC L3-E turntable. It has 680 R coils. I am going to try to fashion a version of this with EQ by ear. Being direct drive it has moments when it puts posh turntables to shame ( Sinatra ). If I can blend in a less coarse sound it might be OK. Like putting cocoa in lower grade real coffee. Hope this is interesting and not a diversion?

This idea for this came from a 1960's Tobey and Dinsdale circuit that I seem not to be able too find on the net. I think they realised people wanted a universal concept where the 75 uS was not related to the cartridge impedance. The one I see is much like the Leak amps.

C5rM1gs.jpg
 
Status
Not open for further replies.
Top Bottom