Page 5 of 25

Re: Flat Phono Preamp Based on John's P10 and 2SK389

Posted: Wed Sep 07, 2011 2:22 pm
by mediatechnology
Thanks, I think I have a handful that Al Burdick gave me but I may still take you up on it.

Thought it might be interesting to compare both the bipolar and FET versions of the P10. WRT to the 2SB737 used in the P10, what was the difference between the bipolar MC version and the MM FET version? How were the bipolar versions biased? Was there a link on the non-inverting inputs of the TL07X to a negative bias voltage? I don't think the FET sch shows anything like that.

Re: Flat Phono Preamp Based on John's P10 and 2SK389

Posted: Wed Sep 07, 2011 4:14 pm
by JR.
mediatechnology wrote:Thanks, I think I have a handful that Al Burdick gave me but I may still take you up on it.

Thought it might be interesting to compare both the bipolar and FET versions of the P10. WRT to the 2SB737 used in the P10, what was the difference between the bipolar MC version and the MM FET version? How were the bipolar versions biased? Was there a link on the non-inverting inputs of the TL07X to a negative bias voltage? I don't think the FET sch shows anything like that.
The bipolar version of P-10 is for MC carts so I wouldn't bother unless that is what you are using. There are surely better bipolar devices for MM cart that 737, but why bother when you can use JFETs?

The operating current for bipolar input devices is provided through the 3k feedback resistors. The tl07x opamps were biased up at +V/2 so roughly 2mA of input device class A current. The 2sb737s are PNP devices so collector loads go to -15v and cap to opamp input is reversed. it would have been neater and simpler to use the NPN 2sd786 but the low noise PNP was slightly lower noise and I was always chasing every last drop of performance. I couldn't find a low noise P channel JFET and the old 5103 is not very good by more recent standards.

edit- other changes were 15 ohm gain resistor for MC appropriate gain, and 100 ohm input termination resistors again for MC carts. /edit

JR

Re: Flat Phono Preamp Based on John's P10 and 2SK389

Posted: Thu Sep 08, 2011 5:38 am
by mediatechnology
There are surely better bipolar devices for MM cart that 737, but why bother when you can use JFETs?
To prove a point. ;)

Forgot that the 2SB737 was PNP and that everything needed to be turned upside down until I revisited Burdick's schematic.

What I was thinking about doing was comparing the noise performance of typical NPN BJT vs. NFET using the exact same test circuit, a dummy cart and dialing bias in by varying the DC at the non-inverting inputs. Would then be curious to see what the BJT input bias current was.

Re: Flat Phono Preamp Based on John's P10 and 2SK389

Posted: Thu Sep 08, 2011 10:09 am
by JR.
If I were trying to design for MM and bipolar, the play is to look at the plots if constant NF vs current, on device data sheets, but one catch with MM carts is the R+L, so the optimal tweak for LF range where R is dominant will differ from HF range where L is dominant.

While one would be inclined to do the math based on noise bandwidth, perhaps with some fletcher-munson correction applied, I would probably use the math as starting point, then listen to the noise floor, while tweaking.

The optimal device for MM is not likely to be one of these low Rbb mic or MC devices. Just off the top of my head 2n5088 or something like that? If they still make them?

JR

Re: Flat Phono Preamp Based on John's P10 and 2SK389

Posted: Thu Sep 08, 2011 1:00 pm
by mediatechnology
Yep, 2N5088, 2SC1815 something like that.

Re: Flat Phono Preamp Based on John's P10 and 2SK389

Posted: Thu Sep 08, 2011 7:18 pm
by ricardo
JR. wrote:The optimal device for MM is not likely to be one of these low Rbb mic or MC devices. Just off the top of my head 2n5088 or something like that?
It's not that low rbb devices are not optimum but that they are unnecessary. The L in MM, means you have to optimise noise performance for fairly high Z. You end up with Ic about 20 - 100uA. This is Rnv of 1k25 to 250. What is important is high Rni, hence NE5532 instead of AD797 or better still a LN FET i/p OPA[1]

Old fashion LN transistors were usually just high hfe devices with some care over manufacture for good 1/f noise. Their fairly high rbb didn't matter much when the source in the important 4 - 8kHz is about 20k. High hfe was more important. Probably any small signal BJT with hfe > 400 is good for MM in the correct (low Ic) circuit.

GG Baxandall deals with all this in his WW68 article as well as MEH Chapter 8.

By all means use your Unobtainium 2sd737 etc. You won't get better noise for MM compared to BC184C/214C etc but it won't be worse.

[1] See! I'm not really anti FET. It's just that when you do the maths, you hardly gain 0.5dB better noise compared to NE5532 for MM. And I've gone to serious lengths to investigate Ib through MM including staring through a measuring microsope and test records. :ugeek:

Any recommendations on good FET OPAs? ie at least as good Rnv as NE5532, low offsets and no nasty latching on overload. Experience please. I pretend to read the datasheets. Gotta at least pretend to enter the 21st century :lol:

Re: Flat Phono Preamp Based on John's P10 and 2SK389

Posted: Fri Sep 09, 2011 6:35 am
by mediatechnology
And I've gone to serious lengths to investigate Ib through MM including staring through a measuring microsope and test records.
That's what I'm hung up on WRT using a bipolar input device. I believe your tests, but what I don't understand is how a DC bias on the cartridge that's as big as the signal (or almost) doesn't introduce a mechanical bias (stiffening?) and magnetic core saturation. Although anecdotal, I've heard of SM-57's and input transformers sounding bad after undergoing phantom power faults. I think it may have been Wes Dooley describing how to de-magnetize an SM-57.

I have more 2SK389s than 2SB737s anyway.

Re: Flat Phono Preamp Based on John's P10 and 2SK389

Posted: Fri Sep 09, 2011 1:30 pm
by JR.
Better is always better and JFET is better. I think Ricardo is saying 5534 was good enough back in the day.

Regarding good modern low noise JFETs, I would not be afraid to use one of the uber opamps, and just work around turn-on latch up issues. Knowing the problem exist means they can be overcome, perhaps simpler than designing an uber discrete or hybrid circuit.

Later...


JR

Re: Flat Phono Preamp Based on John's P10 and 2SK389

Posted: Fri Sep 09, 2011 5:02 pm
by ricardo
mediatechnology wrote:I believe your tests, but what I don't understand is how a DC bias on the cartridge that's as big as the signal (or almost) doesn't introduce a mechanical bias (stiffening?) and magnetic core saturation.
Your analogy isn't quite right. (With my pseudo transducer guru hat) Both MC & MM[1] obey the basic Electromagnetic eqns. of transduction.

The device generates a voltage e = BLv. B: flux density L:length of wire in flux v:velocity
It also generates a force f = BLi i: current through the wire.
The BL factor is the same for both and is the link between the electrical & mechanical worlds. You can calculate the mechanical impedances reflected by the electrical impedances and vice versa.

Knowing the velocity sensitivity of the cartridge, you can work out the force due to Ib. I submit that if this is a very small fraction of the Tracking Force, it can be ignored. Core saturation should show up easily in THD tests.

If you do the same exercise on MC, you see that the effects are quite a lot larger for appropriate circuits (cos higher efficiency of MC cartridges) but still only on the verge of being significant. ... but my favourite MC amplifier has a floating battery supply and hence no input current.

I'd still like some recommendations on FET OPAs from John and yus gurus. Though I've done some tweaky circuits in my time, I'm really quite conservative. I don't recommend stuff unless I've tried it. This Millenium, my only new recommendation is MAX410 cos its the only thing this beach bum has managed to try. It's good for a LN battery powered application even though it has evil Ib cancelling.

[1] Moving magnet & Moving Iron by a slightly roundabout route but this BS still holds.

Re: Flat Phono Preamp Based on John's P10 and 2SK389

Posted: Fri Sep 09, 2011 8:25 pm
by mediatechnology
I submit that if this is a very small fraction of the Tracking Force, it can be ignored.
OK, that makes sense.