LZX Patchable Standard Reference Design Schematics

Forum DIY
1

Attached is a PDF with three schematics for recommended input/output buffering circuitry for production modules. Some notes:

  • These work on a +/-5V power supply. For EuroRack, we typically use LDOs like 7805/7905 to drop the +/-12V input rails to +/-5V. There are lots of LDOs and options to choose from for power, but that’s outside the scope of this schematic.
  • Included are examples for 1V in/out buffers at unity gain, 1V to +/-1V and back again for bipolar processing, and 1V to 5V TTL and 5V TTL to 1V for logic signals.
  • All circuits should perform at full bandwidths if careful PCB layout guidelines are followed. They should be as close as possible to their in/out jacks and a 4-6 layer PCB with full reference ground plane is recommended.

LZX Interface Examples RevA.pdf

Hopefully this helps some of you out! These circuits are identical to what you will find in most LZX production modules. I will post some more schematics soon, such as for sync/genlock input and try to do whatever else you request.

9 Likes
2

ADA4851 looks nice, but MSOP! :slight_smile: I’m getting a bit old for pin spacings that small :wink:

I was playing with the MAX439x series in SOIC which (as far as I can tell) comes in a little bit cheaper (in the dual package at least, which I prefer so that I can use the more convenient 0.3" leg width SMT converters). Similar specs otherwise though, I think. What do you think of them Lars?

Thank you for sharing these schematics.

1 Like
3

MAX439x looks like a very similar part that I bet works nicely! ADA4851-4 has been very flexible and performing great for us – it’s a bit less than the MAX4396 in full reel quantities. I’m sure in most practical cases in the LZX system, both parts would perform well.

I wish there were suitable +/-12V parts, but we you really gotta go +/-5V for video these days! LM6172 is just too expensive for an op amp heavy design.

Any other “bread and butter” schematics you’d like to see posted?

1 Like
4

I guess a “canonical” attenuverter input reference would be nice to have?

I’m slightly confused as to why the bypass caps are drawn with resistor symbols, but apart from that this is fantastic.

Any thoughts in the long term for a +/-5V psu connector standard?
I very much feel the need to be euro compatible, but burning more than half the power in a power hungry design in linregs feels wasteful.

1 Like
5

I guess a “canonical” attenuverter input reference would be nice to have?

There are some basic circuits in the Cadet schematics. We’re moving to all active signal paths moving forward, so an attenuverter is a dual discrete VCA with a special buffered driver circuit from the pot. I’m not quite ready to publish that circuit, but I’ll think about it.

I’m slightly confused as to why the bypass caps are drawn with resistor symbols, but apart from that this is fantastic.

Hah, me too! I must have made a drafting error.

Any thoughts in the long term for a +/-5V psu connector standard?
I very much feel the need to be euro compatible, but burning more than half the power in a power hungry design in linregs feels wasteful.

Lots of thoughts. I had considered a 4-pin Molex MicroFIT connector (an actual power connector, unlike ribbon cables.)

We’ve got switching +/-12V to +/-5.5V followed by high grade RF +/-5V LDOs on power entry to some new prototypes.

The question is very complicated – in the end, for us any product for the modular line we release would need to support both +/-5V and +/-12V even if some sort of alternate connector were offered. We’ve thought about various types of “EuroRack to +/-5V” adapters and plugin assemblies, but all the cost engineering ends up pointing back to nominal differences in costs and a big jump in user frustration with dealing with multiple standards.

One approach is the Roland Aira method, which used only the +12V power rail (available on EuroRack supplies) and switch that to +/-5.5V, then drop with LDOs. It is similar to the approaches we’re testing now, but currently we’re trying to source the negative current from the negative rail to maximize the EuroRack power supply.

So it’s most likely we’ll work around the existing standard, which is +/-12V input. The most important priority for me is that the modules perform really well under any conditions (even bad EuroRack power supplies.) If there’s a strong push for it, we could document a +/-5V (or +/-6V, for LDO entry per module, which is ideal) connector. I’d rather do that than see a branching standard emerge.

Let me know your thoughts if you are seriously considering developing modules under the standard!

2 Likes
6

There are some basic circuits in the Cadet schematics. We’re moving to all active signal paths moving forward, so an attenuverter is a dual discrete VCA with a special buffered driver circuit from the pot. I’m not quite ready to publish that circuit, but I’ll think about it.

I’d love to see those if you get there :slight_smile:
would you be willing to share some thoughts as to why you’re making that change and what the tradeoffs are? I’m guessing you’re beating the cost of a 1251 by a wide margin, but i am a bit surprised to see you use it for everything like this.

Let me know your thoughts if you are seriously considering developing modules under the standard!

I’m not planing to go into “the business” anytime soon, but i will be supplementing my cadet with some of my own designs going forward.
The plan is to at the very least open source anything that turns out to work, maybe running some DIY PCBs / partial premount PCBs if there turns out to be interest, if i can figure out distribution, and very much if i have the time and energy.

3 Likes
7

would you be willing to share some thoughts as to why you’re making that change and what the tradeoffs are? I’m guessing you’re beating the cost of a 1251 by a wide margin, but i am a bit surprised to see you use it for everything like this.

Oh, there are tons of advantages to using the LT1251/LT1256. The biggest cost saver is that if you’re using accurate voltage reference circuits, you don’t need any trimmers!! They are laser calibrated in the factory. Trimmers easily can cost $1-$2 each even in quantity, for a nice quality sealed option, and then add another dollar of tech labor, and discrete options with the same performance can quickly match the cost of LT1251/LT1256. Another advantage on the DFM side is much lower component counts. On the design side of things, it’s dead simple and always works, too – whereas more complex discrete circuitry can end up requiring more revision cycles. In 2015 as we launched into Expedition series development it was very important to expand the product line quickly, so LT1251/LT1256 has been a life saver in this regard.

That said, now that we have stabilized and can afford longer R&D and revision cycles, we’re very interested in optimizing the product line as a whole, for testability, performance, and consistency in manufacturing. Not that we have a lot of problems to solve – we just want to continue doing the best job possible. Since most purchasing is done for the product line rather than an individual product, we’re able to hit some quantity breaks we were not able to before (whereas LT1256 qty pricing ends up capping out around 100 pieces.) So we’re auditioning some discrete circuits now, and also have the kind of test gear we need to do some more advanced signal integrity testing of all the modules.

With active attenuverters and signal paths (like using analog switch ICs controlled by mechanical switches) there are numerous signal integrity advantages. One of them is that the control pot voltage can be manipulated to fine tune the range and slope of a control in the way putting the video through a pot directly cannot. For example, in our active attenuverter circuit there is a deadband region in the center, so you can reach a nulling point easier.

10 Likes
8

Was this in the Cadet git page that is undergoing revision? The pdf link is dead :frowning:

9

Looks like it’s here now: lzxdocs/static/pdf/reference-designs/lzx-interface-examples-reva.pdf at master · lzxindustries/lzxdocs · GitHub

1 Like