AMB linestage rebuild

What people are working on at the moment
User avatar
Thermionic Idler
Old Hand
Posts: 1018
Joined: Sun Apr 22, 2012 8:22 pm
Location: Southsea

#16 Re: AMB linestage rebuild

Post by Thermionic Idler »

Ray P wrote: Fri Jan 28, 2022 4:13 pm I think that has spoiled it Dave.

IMO, the push buttons need to be stacked on the left (a la Quad), move the disply to the right where your buttons are horizontally located and size the cutout etc. for a slightly larger volume knob.
I spent ages pondering that this afternoon - whether to do it a la Quad. The dilemma came down to how the buttons would be labelled (and I don't want to leave them unlabelled as I'd then get the same old questions about which source was which). The 'outer' front panel is powdercoated black from Schaeffer, and as far as I know, they cannot print onto that surface.

With the buttons in their current location below the inner perspex, there can be backlit etched lettering above each one and an LED indicating which is active - readable in low-light conditions. The effect I'm going for is kinda like those old FM tuners that had the backlit dial all along the top. Certainly if I moved the display and stacked the buttons, it would more closely resemble the FM4, but I didn't want to clone the design, just use it as a starting point. I looked at the more recent Quad stuff (QC24) and asked myself what the designers would have done had the engineers told them the buttons needed to be located elsewhere, and tried to come up with an answer.

I'll ponder further though, and all ideas are welcome.

EDIT - I suppose I could have the source names stacked on the left side of the perspex, so:

disc
server
av
aux

corresponding to the vertical button stack, then an indicator LED to the left of each word to indicate which one was active.

That might be better. Easier to implement too, as the pcb won't have to share space with the perspex.

WRT the size of the volume knob, that style doesn't come any larger than 40mm diameter.

What do we think?
Deck: Garrard 301 - Audio Origami PU7 - NW Analogue DH3S
Phono: Pete Millett LR, Lundahl SUT, AMB σ22 PSU
Linestage: Bruno Putzeys Balanced Pre + Hans Polak relay mod, LCDuino control
Power amps: 300B PP 'Symmetric Reichert'
Speakers: FF225WK Big Mets
User avatar
Thermionic Idler
Old Hand
Posts: 1018
Joined: Sun Apr 22, 2012 8:22 pm
Location: Southsea

#17 Re: AMB linestage rebuild

Post by Thermionic Idler »

I'll come back to the front panel design later.

In order to use my existing LCDuino infrastructure successfully, I needed to confirm my understanding of how they implemented their stepped relay attenuator, and then check that the same control system would function properly with the new circuit. I know that people cleverer than me could probably have worked it out but I wanted to prove it in Spice. Here's how I did it.

Based on a six relay control, the first thing I had to do was set up the relay triggers for a binary count from 000000 to 111111, to simulate the output from the control system. The expected result would be the volume either decreasing from max to min, or the reverse. Easy to do, just a matter of doubling the period from least to most significant bit:

Image

The output voltages were made slightly different to make it easier to read the sim results, this confirmed I'd set up the triggers correctly. You can see that at 0ms, all triggers are low, and at 63ms all are high:

Image

So, now we go to AMB's resistor value calculator, get the values for 6 relays, and copy the delta1 attenuator board schematic into Spice, making sure that Set and Reset are the right way round as that determines the volume direction. (LCDuino uses latching relays).

Image

Run the sim, and looking at the voltage across RT0, we see a logarithmic volume descent:

Image

So what we want to do now, is hang the new circuit around the same 6 relays and check that we get the same result. To recap, Bruno Putzey's circuit is at the foot of this article and here's Hans Polak's switched relay modification. He since came up with a subsequent 6 relay design based on the same principles, and that is what I've decided to go ahead with as I would still like to retain a few dB's of gain at the preamp stage. The previous configuration I was investigating would have resulted in a unity gain stage.

So we first copy the Hans Polak schematic into Spice using the same relay set, followed by the Bruno Putzey linestage (I didn't bother with the 'dc servo' bit). Notice how much lower in value all the resistors are compared with a conventional attenuator:

Image

And we check the output...

Image

And there is a match with the AMB attenuator, confirming that if the same latching relays are used, connected per the Spice sim, and interfaced to LCDuino correctly, everything should work perfectly.

When I originally modelled the Hans Polak circuit, I had Set and Reset the wrong way round, so I got a volume increase instead of a decrease.

I'll be laying out my own PCB for this as I want to integrate the input switching, the volume control, and the linestage circuit all on the same PCB, which I'm hoping I can get narrow enough to mount vertically to the back panel with the XLR sockets soldered directly to the board. If I can pull that off, there will be no board-to-board connectors in the signal path at all. Contrast that with my current arrangement:

XLR to I/O selector board - one connector;
I/O selector to volume control - one connector;
volume control to linestage input - two connectors;
linestage output back to I/O selector - two connectors;
I/O selector to output XLR - one connector.
7 connectors total!
Last edited by Thermionic Idler on Sun Jan 30, 2022 12:08 pm, edited 3 times in total.
Deck: Garrard 301 - Audio Origami PU7 - NW Analogue DH3S
Phono: Pete Millett LR, Lundahl SUT, AMB σ22 PSU
Linestage: Bruno Putzeys Balanced Pre + Hans Polak relay mod, LCDuino control
Power amps: 300B PP 'Symmetric Reichert'
Speakers: FF225WK Big Mets
User avatar
Nick
Site Admin
Posts: 15706
Joined: Sun May 06, 2007 10:20 am
Location: West Yorkshire

#18 Re: AMB linestage rebuild

Post by Nick »

Sounds like this

https://en.wikipedia.org/wiki/Logarithm ... tor_ladder

Plus side, less relays, down side more relays in series at once,
Whenever an honest man discovers that he's mistaken, he will either cease to be mistaken or he will cease to be honest.
User avatar
Thermionic Idler
Old Hand
Posts: 1018
Joined: Sun Apr 22, 2012 8:22 pm
Location: Southsea

#19 Re: AMB linestage rebuild

Post by Thermionic Idler »

Nick wrote: Sun Jan 30, 2022 10:12 am Sounds like this

https://en.wikipedia.org/wiki/Logarithm ... tor_ladder

Plus side, less relays, down side more relays in series at once,
Looks pretty similar, yep.

My biggest worry with this is accidentally getting the set / reset logic reversed, so I think when I make up the PCB's, I'm going to have separate identical connectors for each relay bank's send and return - that way I can change polarity just by swapping the leads around rather than having to remake the whole damn board.
Deck: Garrard 301 - Audio Origami PU7 - NW Analogue DH3S
Phono: Pete Millett LR, Lundahl SUT, AMB σ22 PSU
Linestage: Bruno Putzeys Balanced Pre + Hans Polak relay mod, LCDuino control
Power amps: 300B PP 'Symmetric Reichert'
Speakers: FF225WK Big Mets
User avatar
ed
retired
Posts: 5384
Joined: Thu Jun 21, 2007 4:01 pm
Location: yorkshire
Contact:

#20 Re: AMB linestage rebuild

Post by ed »

once again, not sure if this is relevent but....

Back in 2012 I built a succession of ‘passive pre amps’ using atmel attiny461 and mcp4241 digi pots with IR control and rotary encoders. They were partially successful but sq was never top notch. It occurred to me that the digi pots may have improved since those days and there maybe merit in revisiting. In regard to the mcp chips I could never work out what the resistor config was although I know there were 120 odd resistors in each pot.
pp1.jpg
pp1.jpg (78.4 KiB) Viewed 8059 times
pp1inbox.jpg
pp1inbox.jpg (55.92 KiB) Viewed 8059 times
There's nowhere you can be that isn't where you're meant to be
User avatar
Thermionic Idler
Old Hand
Posts: 1018
Joined: Sun Apr 22, 2012 8:22 pm
Location: Southsea

#21 Re: AMB linestage rebuild

Post by Thermionic Idler »

I investigated an implementation that used one of those 'pots on a chip' but decided to stick with the relays, partly because I'd have had to do a lot more software hacking if I did change to a digipot.

Yes, there is the clickety-clack when changing volume but we've actually grown to like that, as it gives us a small audible cue that the linestage is responding to the remote and actually doing something. With the new circuit it'll be a little quieter as only half the number of relays will be switching.
Deck: Garrard 301 - Audio Origami PU7 - NW Analogue DH3S
Phono: Pete Millett LR, Lundahl SUT, AMB σ22 PSU
Linestage: Bruno Putzeys Balanced Pre + Hans Polak relay mod, LCDuino control
Power amps: 300B PP 'Symmetric Reichert'
Speakers: FF225WK Big Mets
User avatar
Nick
Site Admin
Posts: 15706
Joined: Sun May 06, 2007 10:20 am
Location: West Yorkshire

#22 Re: AMB linestage rebuild

Post by Nick »

Thermionic Idler wrote: Sun Jan 30, 2022 10:49 am
Nick wrote: Sun Jan 30, 2022 10:12 am Sounds like this

https://en.wikipedia.org/wiki/Logarithm ... tor_ladder

Plus side, less relays, down side more relays in series at once,
Looks pretty similar, yep.

My biggest worry with this is accidentally getting the set / reset logic reversed, so I think when I make up the PCB's, I'm going to have separate identical connectors for each relay bank's send and return - that way I can change polarity just by swapping the leads around rather than having to remake the whole damn board.
Is that not just a fix in software thing?
Whenever an honest man discovers that he's mistaken, he will either cease to be mistaken or he will cease to be honest.
User avatar
Nick
Site Admin
Posts: 15706
Joined: Sun May 06, 2007 10:20 am
Location: West Yorkshire

#23 Re: AMB linestage rebuild

Post by Nick »

One of the never got round to things I wanted to try was a 4 quadrant multiplying DAC as a volume control.
Whenever an honest man discovers that he's mistaken, he will either cease to be mistaken or he will cease to be honest.
User avatar
Thermionic Idler
Old Hand
Posts: 1018
Joined: Sun Apr 22, 2012 8:22 pm
Location: Southsea

#24 Re: AMB linestage rebuild

Post by Thermionic Idler »

Nick wrote: Sun Jan 30, 2022 11:35 am
Thermionic Idler wrote: Sun Jan 30, 2022 10:49 am
Nick wrote: Sun Jan 30, 2022 10:12 am Sounds like this

https://en.wikipedia.org/wiki/Logarithm ... tor_ladder

Plus side, less relays, down side more relays in series at once,
Looks pretty similar, yep.

My biggest worry with this is accidentally getting the set / reset logic reversed, so I think when I make up the PCB's, I'm going to have separate identical connectors for each relay bank's send and return - that way I can change polarity just by swapping the leads around rather than having to remake the whole damn board.
Is that not just a fix in software thing?
Depends on how well the software has been structured doesn't it - I've tinkered around the edges of it but haven't delved deep enough to find what actually does the grunt work of switching the relays - all I've found so far are calls to "send_byte_to_volume_board". I'll take a look but I suspect swapping a pair of connectors around would be easier.
Deck: Garrard 301 - Audio Origami PU7 - NW Analogue DH3S
Phono: Pete Millett LR, Lundahl SUT, AMB σ22 PSU
Linestage: Bruno Putzeys Balanced Pre + Hans Polak relay mod, LCDuino control
Power amps: 300B PP 'Symmetric Reichert'
Speakers: FF225WK Big Mets
User avatar
Thermionic Idler
Old Hand
Posts: 1018
Joined: Sun Apr 22, 2012 8:22 pm
Location: Southsea

#25 Re: AMB linestage rebuild

Post by Thermionic Idler »

Nick, I have another question for you if I may. I had a bit of a lightbulb moment this morning.

So here's the schematic for the Bruno Putzey preamp circuit, the subject of this thread:

Image

And here's the schematic for AMB's alpha 24 that I currently have:

Image

They are quite similar to each other with that first instrumentation amplifier stage.

Now, my understanding is that in the Bruno circuit, U3 is there to remove DC by acting as a very low pass filter, with its output being subtracted from the signal. And U2/7A following it is intended to be the difference amplifier that removes common mode noise.

Looking at the alpha24 schematic, it looks to me as though Ti Kan has accomplished pretty much the same thing rather more elegantly with the OPA1632 op-amp, doing both jobs in one stage. I think I can replace U3 and U2/7A in Bruno's circuit with a similarly configured OPA1632, doing away with that TL072 filter circuit and its electrolytic caps. That can then feed into the U2/7B volume controller stage.

Am I right or are there any gotchas I need to be aware of?

If it helps, comments from the designers themselves: Bruno circuit

and
Ti Kan wrote:Of particular interest is the Vocm pin on the OPA1632. By connecting it to ground, the differential output common mode voltage is now referenced to ground. It shifts the output common mode "zero" reference to ground potential (even when there is a large input common mode DC offset), and thus avoids the use of coupling capacitors. Due to tight-matching of devices within the opamps, and due to balanced effective resistances on all opamp input pins, it doesn't even need a DC servo circuit to maintain very low offsets. This functionality is particularly useful when the α24 is used as a DAC analog output stage, where the DAC chip has a large common-mode output DC offset (relative to ground) due to its single-rail power supply.
Many thanks.
Deck: Garrard 301 - Audio Origami PU7 - NW Analogue DH3S
Phono: Pete Millett LR, Lundahl SUT, AMB σ22 PSU
Linestage: Bruno Putzeys Balanced Pre + Hans Polak relay mod, LCDuino control
Power amps: 300B PP 'Symmetric Reichert'
Speakers: FF225WK Big Mets
User avatar
Nick
Site Admin
Posts: 15706
Joined: Sun May 06, 2007 10:20 am
Location: West Yorkshire

#26 Re: AMB linestage rebuild

Post by Nick »

The bit you talk of looks like a DC servo, so I would not worry about the caps used, the signal is not going through them. I guess its there to remove and DC offset created by the preceding op amps in as you say a instrument amp setup. C4/C5 is likely to achieve similar in the AMB circuit,
Whenever an honest man discovers that he's mistaken, he will either cease to be mistaken or he will cease to be honest.
User avatar
Thermionic Idler
Old Hand
Posts: 1018
Joined: Sun Apr 22, 2012 8:22 pm
Location: Southsea

#27 Re: AMB linestage rebuild

Post by Thermionic Idler »

Nick wrote: Mon Jan 31, 2022 12:34 pm The bit you talk of looks like a DC servo, so I would not worry about the caps used, the signal is not going through them. I guess its there to remove and DC offset created by the preceding op amps in as you say a instrument amp setup. C4/C5 is likely to achieve similar in the AMB circuit,
Which explains the presence of the cheaper TL072 op-amp as well rather than the LM4562 specified for elsewhere. Guess I don't need to worry about that either.

Aside from substituting the OPA1612 in place of the LM4562 (as I'm familiar with how good it sounds in the alpha24), I think I'll just leave the circuit as is.
Deck: Garrard 301 - Audio Origami PU7 - NW Analogue DH3S
Phono: Pete Millett LR, Lundahl SUT, AMB σ22 PSU
Linestage: Bruno Putzeys Balanced Pre + Hans Polak relay mod, LCDuino control
Power amps: 300B PP 'Symmetric Reichert'
Speakers: FF225WK Big Mets
User avatar
Thermionic Idler
Old Hand
Posts: 1018
Joined: Sun Apr 22, 2012 8:22 pm
Location: Southsea

#28 Re: AMB linestage rebuild

Post by Thermionic Idler »

OK, I've completed the schematic capture in Kicad for the linestage PCB. This incorporates Bruno's linestage circuit, Han's step relay volume control mod and input selection, the idea being that it all goes onto one circuit board and gets driven by the AMB LCDuino control solution.

Link to a PDF of the schematic.

Tomorrow, I export the netlist and find out if I can fit everything onto that single PCB. Can I pull this off? Watch this space...
Deck: Garrard 301 - Audio Origami PU7 - NW Analogue DH3S
Phono: Pete Millett LR, Lundahl SUT, AMB σ22 PSU
Linestage: Bruno Putzeys Balanced Pre + Hans Polak relay mod, LCDuino control
Power amps: 300B PP 'Symmetric Reichert'
Speakers: FF225WK Big Mets
User avatar
Thermionic Idler
Old Hand
Posts: 1018
Joined: Sun Apr 22, 2012 8:22 pm
Location: Southsea

#29 Re: AMB linestage rebuild

Post by Thermionic Idler »

Success! I started a new job at the beginning of February which means my hobby bandwidth has decreased somewhat. However, we now have an integrated rear panel-mounted PCB. I challenge you to find a shorter signal path (passives notwithstanding). The fill areas are outlined so all the tracks are visible.

Image

It became clear early on that in no way was I going to be able to do this with 2 layers - it would have to be a 4 layer PCB. The two inner layers route the set / reset signals to the relay coils, and also have positive / negative power planes. Audio signals travel on the outer two layers, which also have ground planes around the main circuit. The planes are all separated for left and right channels, with two separate power connectors. That gives me the option of 'monoblocking' the circuits with separate power supplies if I want to. The split is down the centre so that the DPDT volume control relays can handle both channels. Also the single TL072 specified in Bruno's circuit has been swapped for two TL071's for the same purpose.

Bruno calls for RFI filters on each input socket (although some solutions have only one RFI filter on the other side of the input selection stage). This has had to be done with SMD components as there was no room for through-hole, with the XLR sockets on the other side of the board. Fortunately, JLCPCB can now populate SMD components for you, so I'm going to have that done for the resistors and capacitors near the input sockets. That'll also be 32 components I don't have to spend time hand-soldering. The op-amps are SOIC-8, but I have hand-soldered those before (when making the alpha-24 boards), so I'll do those myself.

The parts layout follows Bruno's recommendations, they have been placed in pairs, with the tracks carrying differential signals placed close together where possible.

I still need to tidy up the silkscreen labels and place some mounting holes, but overall I'm pretty pleased with this. The jury is still out as to whether I'll make up a second power supply, it depends on how much space there is in the chassis.

The next PCB I need to lay out is a clone of the driver sections of AMB's delta1/delta2 boards. Provided I get that right, it will interface to the existing LCDuino board via I2C, and the software should then be able to drive it all with no changes, although I'll still need to apply mods if I want front panel source selection buttons.

Here's Kicad's 3D rendition of the PCB:

Image
Deck: Garrard 301 - Audio Origami PU7 - NW Analogue DH3S
Phono: Pete Millett LR, Lundahl SUT, AMB σ22 PSU
Linestage: Bruno Putzeys Balanced Pre + Hans Polak relay mod, LCDuino control
Power amps: 300B PP 'Symmetric Reichert'
Speakers: FF225WK Big Mets
User avatar
Thermionic Idler
Old Hand
Posts: 1018
Joined: Sun Apr 22, 2012 8:22 pm
Location: Southsea

#30 Re: AMB linestage rebuild

Post by Thermionic Idler »

Ray P wrote: Fri Jan 28, 2022 4:13 pm I think that has spoiled it Dave.

IMO, the push buttons need to be stacked on the left (a la Quad), move the disply to the right where your buttons are horizontally located and size the cutout etc. for a slightly larger volume knob.
Ray - you were right. I've spent weeks pondering how to position those buttons in an attractive way and still have backlit orange labelling for them.

I woke up about 4:30 this morning and realised how it could be done:

Image

The original Quads had 8 buttons, I only need 4 - but why not use the other 'cutouts' for source labelling with backlit perspex behind, and have a single stack of LEDs to indicate the source up the middle?

Uses these nice oval button caps.

Still not entirely happy with the power button but for the source switching, I think that's nailed it.
Deck: Garrard 301 - Audio Origami PU7 - NW Analogue DH3S
Phono: Pete Millett LR, Lundahl SUT, AMB σ22 PSU
Linestage: Bruno Putzeys Balanced Pre + Hans Polak relay mod, LCDuino control
Power amps: 300B PP 'Symmetric Reichert'
Speakers: FF225WK Big Mets
Post Reply