Schematic for Pro-ject phono box

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
mzperx said:
I would share my symmetrical input version (with INA217). It was made on request of my friend. He his very happy with that...

I'm very curious, really, but, unfortunately I'm unable to download the attachement - I'm getting zero length file. Could you up it somewhere else - on any file hosting or imagehosting - e.g.

Code:
http://www.imagehut.eu/

:confused: :confused: :confused:
 
Thanx, I did it.

Well - quite nothing to add. It seems to be intended for the MM pickup. For the MC the R6 could be lowered. The impedances around U2 could be higher (particularly R8 bigger, C7 smaller), as the OPA134 has the input noise voltage comparable with 4 kOhm resistance (1,1 uV over the audio band approx - measured by AP) and the noise voltage at the output of the INA 217 is about 20 uV in this circuit. And, in general, the INA 217 is not the optimum choice for the MM preamp due its quite high input noise current - the best solution IMHO is with a low noise FET OP, with unsymmetrical input, indeed.

Nevertheless, this circuit certainly performs very well as it is - the comments above are rather of cosmetic kind.

BTW, By some opinions I met last time, the SSM 2019 "sounds better". But, I can't judge it.
 
Hi,

nice design, which reminds me of my own and which is capable of extremely good performance :)
A few comments on parts and schematic.

The used parts are all lowcost parts. You can improve the design a bit by using similar better parts. The input receiver INA217 can be substituted by the INA163UA (SO-14 package) or the INA103 (DIP-14 or SOL-16).
The INA163 features lower distortion and lower noise than the INA217. The internal feeedback resistances are smaller -->less thermal noise- and the output stage is less noisy.
The INA103 is even better than the INA163, though only slightly (distortion figures) but at double the cost.
If You replace R6 by a series connection of a fixed and a variable resistor (e.g. discrete resistors switched by a DIP-switch or a pot) You can change the gain of the circuit such that it fits the need of Your pickup. I´ve used resistance values that allowed for an input sensitivity range from 0.1mV-5mV for nominal output voltage of 300mV and had excellent results with whatever pickup I used.
This is one great advantageof this circuit topology. It can be used with extreme flexibility. Add some R- and C- variability at the input, add some switch to change between symmetrical or asymmetrical input and You have a full fledged onefitsall solution for Your vinyl probs.
The LR-networks at the INA217´s inputs shall prevent the OP from oscillating with very low source impedances (<10Ohms) but it raises the noise figures remarkably. Most pick-ups feature a higher ohmic value anyway and their inductance will raise the impedance figure with growing frequency even more. So You can omit the input guards. Just think of using them (or test for oscillations), when You want to run an extremly low output MC with this stage. Using the INA103 for more than 12Years now in such a circuit I never had any probs driving the stage from high-output MM´s like Linn K18, to the ever popular mid-output MC Denon DL103 to the low-output enigmatic machine Koetsu Urushi. :) I just used small ferrite beads on the signal cables.

The integrator stage around U3 could be improved with an low voltage noise, low offset OP. The OPA137 is quite high in voltage noise and offset, especially those with suffixes ´A´. Since the OP works in unity gain over most part of the freq range and low impedances/shorts are presented at it´s inputs low voltage noise figures are a plus. Low current noise is just useful below the filters (R9/C6) -3dB point. FET-types are good since they are less prone to offsets induced by largely differing impedances at their inputs which can be as high as ~1megOhm in this case! Low offset figures and offset drift figures are needed, since freqencies below 50Hz are pushed in the second gain stage with a gain of up to ~32 (30dB) and any offsets at the input of this stage are presented at the circuits output 32times higher. The typical +-2.5mV (max. 10mV) input offset of the OPA137xA could lead to an excessively high output offset at the circuits output!
OPA604, OPA134, AD711 and AD8610 are all better here, with my favourite -the AD8610- beeing ~10 to 30times better than the OPA137. Even the lowcost OPA134 is min. 5 times better!
The offset network R9/C6 has a very low crossover frequency (<1Hz). The RIAA/IEC equalization recommendation mentions a low frequency cut of 20Hz (7950µs time constant). You can use the offset circuit not only to correct for offsets but also to equalize correctly after RIAA/IEC. As a positive sideeffect You can lower the values of R9 and/or C6.

The 2120Hz network R8/C7 could be designed with slightly higher ohmic value for R8 to lower the capacitance value. It might be a bit problematic to get values of 100nF with first class capacitors like WIMA FKP2 or Mica types. Since the noise of the complete circuit is mostly dependent on the first gain stage and since the second stage is made up from a FET-type (which have lowest noise figures with impedances in the >5kOhms range) You can safely choose R8 in the 2k-5k range and reduce C7 down to 30nF-10nF without sacrificing much on noise figures.
In theory the preequalization raises the signal level till ultimo. In practise the cutterheads couldn´t withstand ever rising signal levels. So a additional time constant of 3.18µs (~50kHz) was introduced by the cutterhead manufacturer Neumann . This time constant should be corrected for within the equalization network. This can be done with a simple resistor in series with C7 (100-300Ohms). There will be a slight tonal difference between the traditional- and the Neumann-equalization. If You use a simple switch to short this resistance You can try which one suits Your taste best.

The use of an OPA134 as second gain stage OPamp is a very good choice. An AD8610 will bring just rather minor improvements.
All OPs I mentioned are available as DUALs too and will give better results than the OPA137/134 combination. My favourites here are the AD8620, followed by the OPA2134 and OPA2604. The use of DUAL-OPs eases the PCB layout and is of no bad at all in this case.

For the same reason of capacitor choice You could think of designing the equalization network R10/11/12/C8 with higher resistance- and smaller capacitance values.

The 100Ohm R7 is used to guard the output of the OP from large capacitive loads. It is useful when the circuit is connected to the next stage (preamp input or poti) by a long cable or a highly capacitive input. If the circuit is built into an preamp wit its short cable lenghts You can omit R7. Modern OP amps can safely drive such loads. ;)

jauu
Calvin
 
bobolix, Calvin thanks for a valueadded comments!

This preamp was designed for high output MC.
DC resistance of the affected cartridge is 95 Ohm, so I think that it is well matched as possible to the actual input stage (INA217).

In the 1st version LR-network was fully omited, but due to oscillating I have tried to apply 330Ohm ferrite bead. It could help a lot, but it didn't kill tendency to oscillate, thereby I took the recommended stabilization network.

Availabiliy of FKP2 capacitor is really only up to 22nF from my sources, so I will recalculate resistor values.

Calvin, since your post I don't like the applied dc servo solution anymore...

Is it a good way to eliminate dc error at the circuit ouput if the feedback resistor of dc servo stage would be conncted to output of 2nd amp instead of output of 1st?

Or much more recommended to use inverting circuit in position of 2nd stage with local dc servo loop???

How would you update the circuit in order to be capable of balanced and single ended outputs (both together)?
With adding of DRV134?

thanks in advance
Zoltán
 
Hi,

Yup, You´re right. This circuit is optimized for mid- to high-output MCs.

You got oscillation with an high-output MC? Well that shouldn´t happen. I used the INA103 and had never any probs. I´ll try the INA163 though with my next project probabely in January.

I use 10nF Mica 1% and parallel those (2 or 3). This reduces tolerances even further. Since the sound of a phonostage is mainly due to the precision of the equalization I usually simulate the components to an optimum value and invest some bucks on highquality low tolerance parts ;) I´m quite confident with the results (usually less than 0,1dB).

The DC-servo works very good as it is. The first stage needs some measurement against DC. Since You correct for DC in the second stage of the input device You don´t have to worry about coupling caps at the very input of the circuit (a lot of MC-headamps push the poor little signal through big nasty electrolytics). I tried to connect the DC-servo between the circuit´s output and the reference Point of the INA but this gave always stability problems. So I stuck to the original solution. Using OPA2604 for the servo and the second stage amp, I never had problems regarding DC at the output.
There are two measures to cope with the problem if DC occured
- You could add an DC decoupling cap right at the output
- You could change the DC-gain of the second stage to a value of 1 by adding a cap in series with R10
The first measure gives You relly no DC at the output and is the safest method for Your system, but it may degrade sound
With the second measure a negligible small -and safe- part of DC is still there, but You can design the values such that the 1st order equalization below the 20Hz RIAA/IEC (7950µsec) by servo No. 1 is assisted to a second order (-12dB/oct) character.
Anyway, I prefer the simple solution as it is shown in the schematic.

I had an order for a standalone vinyl solution by a former friend which offered the following features:
- symmetrical and asymmetrical Phono-inputs
- switchable input resistances and capacitances
- switchable gain to accomodate from low-output MCs to high output MMs
- one symmetrical and one asymmetrical highlevel-input.
- highest quality volume control with switched discrete resistors (64 steps)
- one symmetrical and one asymmetrical output
- IR controlled Volume, Channel and on/off/mute
- manual control by digital encoder (one knob design) and 2 digit display
- 2 SMPS, one for +-15V audio and one for the Volume controller
- casing routed from a massive aluminium block with 1/2 HiFi width (22cm)

The Phonostage and the two highlevel inputs are switched by the volume controller PCB and the signal is routed through the volume resistor cascade. Then it is buffered by two parallel working buffers, one of which giving an asymmetrical output and the second a symmetrical output. You can use the DRV134/135 or SSM 2142 for highquality, easy and space- and part saving solutions. For the asymmetrical output I use a circuit that is capable of driving large capacitive loads and uses a OP together with an discrete or integrated buffer (e.g. BUF634)

There are two PCBs, one for the volume controller and one for the complete audio circuits. The audio PCB is a doublesided layout which makes intensive use of SMD parts (I used only parts in throughhole technology where they provided for higher quality than SMD, mainly in the EQ-circuit and powersupply)

Recently I found out that Clearaudio offers a quite similar stage (named Balanced something xxx) which uses a similar topology, but with considerable less flexibility and comfort and component quality for a mere 3.000€. I found the measurements -though all in all on the good side- below whats possible with this circuit.

jauu
Calvin
 
:D :D :D
The ClearAudio mentioned above is based on my older design, which I made for one czech company, which manufactures several products marked as Clearaudio (including the engraved front panel), hehe. The measurement results are not very serious, really, although the device performs quite well - but, unfortunately, although we made our own measurements, we have no influence on the marketing policy of ClearAudio.

The original design was with SSM2017 and after discontinuing production it has been redesigned for the INA163. Theoretically, it exists true replacement for SSM2017 made by THAT, but it is very hard to obtain. Now I'm finishing a fully symmetrical design using INA 103.

Some notes more: The HF instability at open input is caused by capacitive coupling between input and output usually. The remedy is rather capacitive shunting between inputs than using of LR decoupling.

And why fear the DC servos ? The OPA177 (and similars) is a very good choice for this purposes. The output noise of the DC servo presents no problem and it can be easily filtered out.

The second stage need not to be FET, if some DC servo is used with sensing of the end output. The AD5671 may be the optimum compromise.
 
arnoldc said:
hi bobolix! Did you design the Clearaudio Basic Symmetry as well? This is their MC-only phonostage with RCA input and very small 5-pin connector with cartridge auto loading.

I was wondering if I can rip it up and make improvements? Although my warranty says I still have to wait 4 more years before it expires :D

Yep, it's my child, except the input RCA - this is a nonsense degrading the symmetry, indeed. There should be the 5pin only, but it would be incompatible with usual connections, so this ... hmmm ... emergency solution. You can make any modification in order to get the input symmetry back, but, about warranty, it is at your risk. Concerning other improvements, you can consult them with me (recommended).
 
bobolix said:


Yep, it's my child, except the input RCA - this is a nonsense degrading the symmetry, indeed. There should be the 5pin only, but it would be incompatible with usual connections, so this ... hmmm ... emergency solution. You can make any modification in order to get the input symmetry back, but, about warranty, it is at your risk. Concerning other improvements, you can consult them with me (recommended).

I'd take you up on that offer! Should I send you and email and we'll take this offline?
 
Hi bobolix and Calvin,

could you please look at the attached version and give any further commnets?

I have changed the input termination circuit according to Federico Paoletti's schematic. I feel that it's matched better to RCA inputs.
Impadances are also changed in equalization circuits.
In order to keep dc error at the output low as possible, gain of 2nd stage was decreased. As per my calculation it can be max 40mV in worst case.
And balanced output is introduced too.

regards
Zoltán
 

Attachments

  • ina217_riaa_v1.3.pdf
    7.7 KB · Views: 907
Hi,

so far the circuit does what it is expected to do in a simulation. The lower mids are slightly elevated (+0,5dB@200Hz) and -1dB-freqs at 36Hz and 22kHz, resp. 75kHz (switch at 50kHz position). Gain @1kHz is ~40dB, giving 775mV Output voltage for 5mV of Input voltage.
Reducing R2 to 10Ohms will give a gain of 60dB and 775mV Out for 0.5mV In. The second stage doesn´t support the gain above 1kHz so the gain of the first stage (@1kHz) is basically the gain of the whole stage. While the INAs of BB work very good in a gain window of 50-500 (best with a gain of 100) and this is what You need for MM-use, I prefer to give the second stage some gain too (around 10dB) to supply for enough gain even for low output MCs without stressing the INAs with too high gains.
I find the passive EQ-circuit quite clever. While the classical RC-circuit (with additional switchable resistor connected to ground) gives slight deviations (0.1dB) between Standard- and Neumann-EQ (50kHz), the design in the above PDF gives identical frequency responses independent of the switch´s position.

What do You expect from R3/C7 at the Input? It earthes the RCA-ground but leaves only asymmetrical signal handling left. In this case a symmetrical input device like the INA is of no better use than a classical OP. Indeed are the noise figures higher than with an OP which states the same noise figures in the datasheet(!) since there are two equal inputs connected in series noisewise. With uncorrellated noise signals the actual noise figures of an instrumentation amp has to be multiplied with SQR2 to give the correct numbers. The use of an INA makes really sense when the signal is fed to the circuit in a symmetrical fashion! If You just want to use MMs -which mostly are no symmetrical generators because of internal grounding of the coils- You´d be better of with devices like LT1028 or AD797!

jauu
Calvin
 
Hi (Calvin),

I intended to ensure lower impedance for shielding of RCA cable by adding C7/R3. I knew that this kind of solution will degrade capability of INA, but I thought that it's a best what I can do in case of RCA inputs.
Of course the optimal solution would be to apply XLR connectors for phone cartridge inputs where the shielding is conected independently from signal wires.
In practice it can be achieved with a simple connector change on the end of phono cable so I have decided that I will populate only XLR type input connectors on PCB. Also the RCA asym. output connector is omitted.

At version 1.3 the frequency response deviation was due to my faulty calculations at 2nd stage. I applied new EQ circiut where componenet values can be easier to determine and and 3x gain was added to this stage also.

I have rechecked/simulated the new response with TINA-TI (and with help of inverse RIAA network from Lipshitz), and it seems that the overall accuracy is in +/- 0,1dB tolerance now. (Unfortunately this version allows to simulate only 2 OPAs at a time. So I could simulate the fully preamp only partially or in more steps.)

There are few new features introduced as selectabe input impedance and gain seetings.

Please find the SCH of modified preamp and please don't hasitate to reflect on design!

Happy New Year! And lot of succesful projects!

regards
Zoltán
 

Attachments

  • ina103-riaa_v1.4.pdf
    11.2 KB · Views: 2,851
Thanks

Hi Calvin,

I would like to say thank you for your helps.
The phono stage is fully completed and it sounds really impressive.

greetings
Zoltan
 

Attachments

  • mc-mm_sym_riaa_lowres.jpg
    mc-mm_sym_riaa_lowres.jpg
    44 KB · Views: 2,238
Hi Zoltán,

a very clean board, congratulations. :cool:

Just like Calvin, I'd also like to know how it sounds and how it compares.
And I've got another couple of questions, if you don't mind... ;)
---

I can see you used Wima FKP2 capacitors in the front end.
In that range and size you can buy MKT, FKP and FKC types (despite mica and ceramics).
Any particular reason you chose the FKP dielectricum (despite the availability in better selection grades)?
---

I can see you attach power to the board by another XLR plug, right? Funny. :)
Why do you go for no local power supply decoupling near the ICs at all?
I assume this detail is just a difference between the posted schematic and the board you made, as the additional caps are easy to spot near the INA, OPA and DRV...
---

What are the ferrite beads doing, and why?
---

Are you aware of the 'pin one problem'? How is noise immunity with your board and your connections, now that you seem to mix signal and power supply ground in the high-gain stage?
---

Thank you for your effort and sorry for all the bothering questions. :D

Cheers,
Sebastian.
 
Hallo Calvin und Sebastian,

First of all I must say that this phono stage isn’t working in my system (I prefer only digital sources), but one of my friend’s quit expensive chain. He has ordered it, because his new 2nd hand ML38 preamp doesn’t contain phono input. In order to satisfy your interest, I have already requested a short description/comparison concerning sound quality from my friend. As soon as I have it I will share with you.

To return to the questions:

The choice of FKP2 was based on Cavin’s suggestion, and in the other hand the easy availability of this WIMA series at local suppliers.

Between the outputs the 5pin XLR male connector can receive the external PSU, because I didn’t want to put the transformator in same case together with the preamp. (And optional the user can use batteries as well…) Currently one modified Jung regulators is used as power source in same size ALUBOS housing then the preamp.

Electrical point of view the ferrite beads don’t play to much role in this low frequency application (increasing impedance of bypass loops), but their usages and lengths allow to increase the surface of ground plane around ICs.

The ‘pin 1 problem’ was absolutely new for me. But I’m sure that it isn’t issue in case of this design because I/we didn’t find any hum & buzz.

And finally please find a picture about a topview… And don't hasitate to tell me your critisms!

Tschau
Zoltán
 

Attachments

  • pcb_topview.jpg
    pcb_topview.jpg
    77.1 KB · Views: 2,264
I spent some time at designing completely new circuitry - in cooperation with my business partner, so I can't publish it, unfortunately. Note it is a MM-MC switchable fully-active-equalized phono preamp with tolerance ±0.2 dB inside the audio band. They were made extensive tests around distortion using Audio Precision System 2712 and it was found, that certain dual opamps exhibit strange behavior concerning odd harmonics. Evidently there originates certain error voltage at internal IC traces of one subsystem and this error is transferred into signal path of the other subsystem of the dual opamp. Particularly the new AD 8599, although being very good concerning noise, is quite susceptible to this kind of distortion (frankly said, they are some hundredths or thousandth percent THD in the game). In contrary, the NE 5532 seems to be very good from this point of view. But, the work is not finished until now...
 
Hi all

Since this thread started on the Pro-Ject Phono Box, if we could come back to it for a moment...

Does anyone have the schematic for the Pro-Ject Phono Box MkII?
This is a whole different animal from the original - MM/MC switchable and TL 071 MC input gainstage.

The only thing it dosen't offer is variable loading, so I would like to make suitable changes for my DL103. This gizmo uses a small, double-sided PCB, so the component placement is not all that obvious.

Any help would be greatly appreciated :D

Jess
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.