Resurrecting a 1980s analogue preamp project

[MarcAL]

They even let me keep the faulty LED's... the Holy LED's.

[Wicked_XR]

Completed the preamp, revamp board and built.....now to put it back in.

[mrbugeyes]

WELL DONE.  CHEERS - M

[thatmetalman]

Hi everyone.

 

Short break away, so refreshed and motivated to get this across the line.

 

All the mute transistor work is complete, as are the component changes per the designer's recommendations from earlier this month:

• 0.1uF mono across op-amp supply rails (not a pair per op-amp, per Doug Self & David Tilbrook recommendations)
• Op-amp output coupling cap value change to 470uF per Doug Self recommendations and tests (approved by David Tilbrook)
• Mute transistors changed for reed relays; with relevant/associated component changes
• Output connectors changed to smaller, low profile versions to save space, and moved towards the front of the board to allow for better rear panel access
• Op-amp components condensed
• +- 15V change to +- 17V per rail for each op-amp, per Doug Self recommendations and tests
• All signal grounds isolated per original design

Next steps

• Next stage consultation with original designer for checks and sign-off
• Breadboard tests of PSU and mute circuitry control

 

 

Edited November 21, 2024 by thatmetalman

[Leinster Lad]

Wow Tim, some great work being done, and with next level technical input from DT & DS as well  ...... AMAZING !!!

 

Very excited about this actually

[Leinster Lad]

Is the -15v track missing that feeds the LED meter power socket ??

 

 

[thatmetalman]

3 hours ago, Leinster Lad said:

Is the -15v track missing that feeds the LED meter power socket ??

 

 

 

Hah! Keen eyes! Bill picked on that as well, as did I, but not until I'd uploaded the images. It has been corrected.

[MarcAL]

Very impressive!

 

I think that I may put my spare original board rebuild on hold and perhaps get one of these if they hopefully become available.

 

Magnificent work Tim.

[thatmetalman]

19 hours ago, MarcAL said:

Very impressive!

 

I think that I may put my spare original board rebuild on hold and perhaps get one of these if they hopefully become available.

 

Magnificent work Tim.

 

Cheers. Boards will available for purchase for those keen for a direct swap-out of the originals.

 

As a phase 2, I am VERY keen to have complete kits produced including chassis. I started the front panel redesigns but I want to focus on the boards first. Coincidently, the original chassis and front panel supplier was located just a few streets away from where I now live here in Sydney - I checked if he was still in business... alas, not. Amusing, too, that the last time I saw him at his workshop in Brookvale (mid-late 90s) he handed me a 5000 preamp front panel as a keepsake, and I threw the bloody thing away 15 years ago because I grew tired packing it up with my stuff as I went interstate and back! Oh dear!

 

There are some nagging issues with obsolete components on the MC board, e.g. LM394CH matched pair. There are workarounds. But I'll discuss with the original designer first and see if there's a moderately simple workaround. A Latvian supplier apparently has an equivalent (AS3194). I did countless hours of research on this. Tilbrook in his original 1981 article went to pains to explain why he chose the 394 in the first place, so we'll see how tough it is to take a different approach.

 

What I have found most interesting in the MB redesign is, that so-called audio-grade electrolytics in the op-amp signal paths, are unnecessary. This follows comprehensive tests by Doug Self (including subjective tests - which I think he said were tantamount to silliness, i.e. test bench results don't lie; everyone has differing levels of hearing ability, and so on). Doug achieved sensational results using "bog-standard" electros; Tilbrook backed up this assertion. I won't comment further because I haven't run A/B tests in the manner that Doug did, and I am happy to take both Self's and Tilbrook's advice on face value - their designs and broader careers, for me at least, speak for themselves. Who am I to argue!

 

Further to the audio-grade cap subject, I spent countless hours hunting down specific values to suit the 5000 preamp. It was painful. Many Nichicon are EOL or depleted (although to be fair, I found the right values but at higher voltages - and physical size - than what the cct calls for; so it is 'doable'). For those still keen to pursue the attraction of Elna, etc., the caps won't break the bank. Most share the same or similar footprint. The original designer repeated a number of times that it's the low noise of the power supply and the isolated signal grounds that are key, not the signal path electros. I'll leave it there because I know some users on this forum believe in those caps 100% - and that's fine of course - I genuinely have no issue with this subject; it's a personal preference. For the record, I ordered Nichicon and Elna caps from various suppliers before I read about the things mentioned. I may or may not use them when loading the boards.

 

Cheers,

 

Tim 

 

 

Edited November 22, 2024 by thatmetalman

[thatmetalman]

Hi folks.

 

I had a long Zoom call with the original designer over the weekend.

 

The key take aways:

• The electrolytic caps in the signal path in/out of each Line and Monitor op-amp were chosen in 1981 because metallised polypropylene (MKP) caps weren't readily available, with electros deemed "acceptable". As MKPs are more readily available today, David felt strongly about swapping the 2u2 and 47uF electros with suitable MKPs if possible.  However, the capacitance ranges were accommodated by high AC voltage ratings resulting in very big footprints. We both struggled during the call to source 2u2 or 47uF MKPs that were cost effective and/or small enough in footprint size. So, a few challenges. Wima make the right values but sourcing them is proving a challenge.
  • David felt that this change is worth the effort - so I will go away and source suitable caps if possible. I will also allow in the PCB design, both MKP and electros for builders to decide. Some of the MKPs we looked at hovered around the $8-11 mark each (one was even A$111!), but to be frank, we're aiming for high-end performance here, and a piece of gear that will last for many, many years - long after the initial outlay of, say, ~A$40 to make a significant improvement.
  • For suitable footprints, SMDs were looked at briefly but David is unhappy about potential for distortion caused by variations in temperature coefficient (I think I got that right!). SMD caps won't be used.
• Reed relays have been approved together with all control cct changes
• Passives on the RCA sockets were approved to be moved to the main board
• For the phono MC and MM boards, David would like to revisit the RIAA components. Both circuits will be revised with the possible inclusion of being able to accommodate different cartridge makes
• The LM394CH super matched pair are long obsolete. Alternative Analogue Devices products were looked at, as well as AS194H from Latvia. David will make the call on this after I source alternatives. This shouldn't be a roadblock.

 

There was also desire to employ integrated LED bargraph displays over individual LEDs. For the time being I will stay on course per the original. A revisit of these boards may be considered for 'phase 2'. But I have the LED/VU boards completed and can release those at least to load up and test.

 

So, a number of immediate design improvements to be employed across the MB and MM/MC boards.

 

Changes to the Monitor and Line amp sections will be prototyped soon and measured, before being compared to an original 1981 build's performance.

 

Cheers,

 

Tim

 

Edited December 2, 2024 by thatmetalman

[MarcAL]

Thanks for the update Tim.

You have done a great deal of work lately and are getting lots of the boxes ticked.

 

One question I have is, is there anyway of reducing the crosstalk within the 5000 as I remember having issues with this back in the day?

Maybe it was partly due to the switches and the quality of the screen cabling used but I did have some issues especially when I was recording to tape.

 

Any thoughts on how to minimise this would be appreciated.

 

Again thanks for the hard work.

 

Marcus

[thatmetalman]

22 hours ago, MarcAL said:

Thanks for the update Tim.

You have done a great deal of work lately and are getting lots of the boxes ticked.

 

One question I have is, is there anyway of reducing the crosstalk within the 5000 as I remember having issues with this back in the day?

Maybe it was partly due to the switches and the quality of the screen cabling used but I did have some issues especially when I was recording to tape.

 

Any thoughts on how to minimise this would be appreciated.

 

Again thanks for the hard work.

 

Marcus

Hi Marcus.

 

First up, I don't think I congratulated you on your 5000 power amp restoration from the other thread; if not, mate seriously, that was super well executed. For the MOSFETs, what are you doing for the 2SJ49 and 2SK134s which I understand are obsolete? Excuse me if you addressed that in the thread and I missed it .

 

Re crosstalk. It could be something to do with the cable impedance (inside and out) but I've emailed David to get it "from the horse's mouth". I'll let you know when he responds.

 

For my rebuild (and ultimate prototype for the preamp) I will use low capacitance cable which is what Jaycar used in their version. It is possible that your kit supplier provided standard cable. Just a guess? My preamp didn't show obvious signs of crosstalk that I recall, but heck it's been a long time! Maybe @Monkeyboi can share his experience with his preamp assuming he has decks/reel to reel connected?

 

I sourced some low capacitance cable in the interim from an eBay seller by the metre but I need to source by the roll for potential kit production. The cable I have appears to be really good and of a diameter perfect for the preamp. Whether it's low capacitance or online BS remains to be seen (heard/measured!).

 

As I write this I'm making the main board mods and pulling a spreadsheet together for David's preferred MKP caps for the signal in/out to/from the op-amps. I'm curious to see what's available after our joint fail quick-look last weekend.

 

Cheers,

 

Tim

 

 

[MarcAL]

1 hour ago, thatmetalman said:

Hi Marcus.

 

First up, I don't think I congratulated you on your 5000 power amp restoration from the other thread; if not, mate seriously, that was super well executed. For the MOSFETs, what are you doing for the 2SJ49 and 2SK134s which I understand are obsolete? Excuse me if you addressed that in the thread and I missed it .

 

Re crosstalk. It could be something to do with the cable impedance (inside and out) but I've emailed David to get it "from the horse's mouth". I'll let you know when he responds.

 

For my rebuild (and ultimate prototype for the preamp) I will use low capacitance cable which is what Jaycar used in their version. It is possible that your kit supplier provided standard cable. Just a guess? My preamp didn't show obvious signs of crosstalk that I recall, but heck it's been a long time! Maybe @Monkeyboi can share his experience with his preamp assuming he has decks/reel to reel connected?

 

I sourced some low capacitance cable in the interim from an eBay seller by the metre but I need to source by the roll for potential kit production. The cable I have appears to be really good and of a diameter perfect for the preamp. Whether it's low capacitance or online BS remains to be seen (heard/measured!).

 

As I write this I'm making the main board mods and pulling a spreadsheet together for David's preferred MKP caps for the signal in/out to/from the op-amps. I'm curious to see what's available after our joint fail quick-look last weekend.

 

Cheers,

 

Tim

 

 

Hi Tim,

 

Thanks for the nice feedback on the Series 5000 power amplifier restoration.

 

I reused the 2SJ49's and 2SK134's from my original amplifier along with the heatsinks, transformers and output choke.

All of the other components were new as I purchased new boards from RCS many years ago.

As far as sourcing Hitachi vintage power mosfets nowadays, I keep an eye out for abandoned amplifiers and do a bit of recycling.

 

I upgraded the supplied screened cable to the low capacitance type that was recommended by Jaycar back in the mid 80's and afterwards still had a margin of crosstalk.

With only a hunch I suspect it is probably from the switches so that may require a *workaround of some sort but this would change everything in this build.

 

* = Another rabbit hole is appearing

 

Regards,

 

Marcus

 

 

[thatmetalman]

On 2/12/2024 at 3:01 PM, MarcAL said:

One question I have is, is there anyway of reducing the crosstalk within the 5000 as I remember having issues with this back in the day?

Maybe it was partly due to the switches and the quality of the screen cabling used but I did have some issues especially when I was recording to tape.

 

Hi Marcus.

 

Response:

I am not aware of any significant cross-talk issue with the 5000. I assume you are referring to inter-channel crosstalk? Or do you mean "bleed-through" from different unselected  sources?

...I am not aware of inter-channel crosstalk issues. This certainly was not a problem in the days of phono and tape inputs, which had much greater inter-channel crosstalk than does the preamp. 

I remember carrying out cross-talk measurements during the development of the preamp but I forget if I included any cross-talk measurements in the article?

 

I could not locate published cross-talk measurements within any of the original ETI articles.

 

Perhaps if you clarify the question bold highlighted above, I can shoot that back.

 

Tim 

[MarcAL]

4 hours ago, thatmetalman said:

 

Hi Marcus.

 

Response:

I am not aware of any significant cross-talk issue with the 5000. I assume you are referring to inter-channel crosstalk? Or do you mean "bleed-through" from different unselected  sources?

...I am not aware of inter-channel crosstalk issues. This certainly was not a problem in the days of phono and tape inputs, which had much greater inter-channel crosstalk than does the preamp. 

I remember carrying out cross-talk measurements during the development of the preamp but I forget if I included any cross-talk measurements in the article?

 

I could not locate published cross-talk measurements within any of the original ETI articles.

 

Perhaps if you clarify the question bold highlighted above, I can shoot that back.

 

Tim 

Hi Tim,

 

I believe that there was some bleed through from different unselected sources.

The main one was from the oscillator which I ended up removing the IC at the time and that fixed that.

 

On the weekend I shall do some more tests to check how much of a problem it is these days as I may not notice that anymore due to my limited HF hearing response.

This may not be an issue for me due to the fact that I can no longer hear tape hiss either.

 

Marcus

[Wicked_XR]

Oscillator shouldn't work unless selected, I would be checking the switches. Tape mute is to stop cross feed....

[Leinster Lad]

Oscillator is  always powered up.

Output fed to High/Tape/Osc switch.  no mute circuit involved.

 

Sounds like either the coax or the switch allowing signal to bleed thru.

[MarcAL]

55 minutes ago, Leinster Lad said:

Oscillator is  always powered up.

Output fed to High/Tape/Osc switch.  no mute circuit involved.

 

Sounds like either the coax or the switch allowing signal to bleed thru.

I am suspecting bleed through on the C&K Lorlin switches.

To be confirmed on the weekend.

[thatmetalman]

G'day.

 

Following David's preference to swap the signal path electros with MKP caps, I thought I'd first see if it was viable given the almost ridiculous physical dimensions of MKPs with the same values (2u2 and 47u). Although 47u MKPs were way too big, 30u devices are available from Wima and TDK (with the 30uF versions employed here after confirming that they were okay to use).

 

The 30uF MKPs are A$12.00 ea, and the 2u2 are A$2.85 ea. Neither destroys the bank but does add to the overall cost. We agreed, however, that the additional cost would be long forgotten after install...!

 

My thoughts on this MKP version. I am more inclined to consider such changes for Tilbrook's AEM6010 redux rather than for the 5000. But, as it all fits, why not? It did take some serious re-jigging to get it all in. A few days.

 

So, as the goal posts keep swaying with the wind, perhaps 2 main boards can be offered - the "MKP board" and the "standard electro board" for simple refits. Or a dual footprint of MKPs and electros for a single board. The height btw will fit a 44mm rack.

 

Happy to take suggestions and thoughts... this latest iteration is certainly "out there".

 

- Tim

 

 

 

[Leinster Lad]

Looking really good Tim.  Can't wait !!

[MarcAL]

That is truly excellent work Tim!!

 

My question is does anyone have a spare or know of anyone who has a spare Series 5000 preamplifier in a reasonable condition that they would want to sell?

 

[thatmetalman]

5 hours ago, MarcAL said:

That is truly excellent work Tim!!

 

My question is does anyone have a spare or know of anyone who has a spare Series 5000 preamplifier in a reasonable condition that they would want to sell?

 

 

I have 2; one is 100% on-spec and okay, the second is parts (main board stripped, LED boards ok, MC/MM boards mostly ok), and chassis/front/rear panels intact. Let me ponder...

[Leinster Lad]

On 21/11/2024 at 8:46 AM, thatmetalman said:

+- 15V change to +- 17V per rail for each op-amp, per Doug Self recommendations and tests

Hi Tim, hope you are well.

Had a spare few minutes to read back over this thread and noticed this change.  +-15V increased to +-17V.

Couldn't find any further info around this ?

 

Also, wondering what is the purpose of  R10 47K in the power supply ?

[thatmetalman]

On 18/12/2024 at 12:58 PM, Leinster Lad said:

Hi Tim, hope you are well.

Had a spare few minutes to read back over this thread and noticed this change.  +-15V increased to +-17V.

Couldn't find any further info around this ?

 

Also, wondering what is the purpose of  R10 47K in the power supply ?

 

Hey David, happy new year mate.

 

The 17v change followed Doug Self's testing with that voltage and the resulting improvements, however, during the last session I had with the original designer, he said that he couldn't see justification and to leave the rails at +-15v. My intention is to bench test against both voltages (which is easy given the LM317/337 adjustable regs).

 

R10 47K I can't locate...?

 

I have the toroidal arriving from Perth this week so should be able to start bench testing the PSU at least. I also have remaining parts for an ultra low noise dual track supply I'm building, so good to have some momentum back.

[Leinster Lad]

Hi Tim and happy new year to you !

 

Be great to hear the results of your bench testing.

 

 

R10 47K  must be from an earlier version of your power supply.