digitally controlled stepped attenuators..??

[Rochey]

[quote author="mcs"]Most of the people here are not trying to make volume controls - they need pots with strange tapers. You could use digipots, but they are not that great (most of them anyway). They have a small number of steps and a very limited voltage range. They also add distortion that the relays don't...
[/quote]

I agree with you in some respects, although without giving too much away (i'm not allowed to) -- you'd be suprised how many of the uber-high-end console manufacturers use these parts for their analogue desks.

The PGA2311 is a -5V to +5v part, whilst the PGA2310 is a -15v to +15v part. Plus and Minus 15v should be more than adequate headroom - no?

if it's just a pot that you want (not a rotary encoder) then some of the MSP430 devices have an onboard 12bit ADC -- allowing you to sample the voltage through the pot, and having a lookup table on teh device... peice of the preverbial p*ss.

Again, I know I sound a bit too pushy on TI's parts, but all I can say in my defence is that it's really hard working with these parts day in day out, and not having great ideas of where they would fit in audio projects! :wink:

I recommend that you take a look at the PGA2310 part (http://focus.ti.com/docs/prod/folders/print/pga2310.html) - I think you'll find that it's quite a sexy part

Cheers

R

P.S -- as for 8051 devices... TI has some data aquisition parts (ADC and 8051 onboard), although to be honest, I'm not so tuned on these parts. You can one of these parts at: http://focus.ti.com/docs/prod/folders/print/msc1200y2.html :sam:

 

[Mbira]

Glad to see this discussion back! Rochey welcome! So the rotary encoders need to have some sort of LED to see where you are right? I thought I saw some rotary encoders somewhere that act more (physically) like pots where they remember their setting when shut off, etc. Does TI or anywhere else have that kind of thing?

Thanks!
Joel

 

[rafafredd]

[quote author="mcs"]But what features do you need on the control board? Buttons, controls, displays etc.
[/quote]

I just want a very simple circuit.

I like the idea of two pushbuttons and a three digit display for 0-256.

That would be nice!

 

[pilo]

Hi everyone!
Rochey : I use pga4311 with MIOS already (and a lot of other TI chips!), and yes you're right it's very easy to work with (even the smd package, big enough for good soldering!). I have some other questions about TI component, can I ask you about those?

For the relay stuff, it's interesting because we can make the pot we want, log, anti-log etc...

I can make a small javascript apps to calculate the value of the resistor, if no body make it before.

 

[Rochey]

wow, what a positive reaction! I'm always paranoid about revealing my employer --- i always think people will start complaining that i'm too biased etc. This is not what I expected! :grin: -thanks

Pilo:
I have some other questions about TI component, can I ask you about those?

Please feel free to ask me questions, although I don't know how often I'll have the answer to hand. If they are questions I can't find the answer to, I'll probably be able to direct you to the PIC (product information center) or the EPIC (European PIC). There's a good team of people there who really have their heads screwed on.

If your going to ask questions, I suggest you do it here in the forum, because there are times when people have the same question etc.
BTW -- I AM NOT DEEPLY TECHNICAL -- there, that's out of the way

Glad to see this discussion back! Rochey welcome! So the rotary encoders need to have some sort of LED to see where you are right? I thought I saw some rotary encoders somewhere that act more (physically) like pots where they remember their setting when shut off, etc. Does TI or anywhere else have that kind of thing?

As far as I know, the majority of encoders just come in a pot like package, without any display, and without any 'memory'.
Don't forget that R.E's just give a 2 bit grey code output. They don't log position in any way. It's kinda like hanging a painting... the rotary encoder is like a friend that's saying "Higher, no lower, lower again etc"... it can only give you direction relative to where it was in the 2 bit code.

For the memory type function that you suggest, I think you would require a multibit output R.E (with say 6 turns with a 16bit grey code). I remember having a look for these at one point, and they cost A FORTUNE! It's signficantly cheapers to use a bit of flash memory (often in small micro's already) to remember the position of the controller. It also means that you can switch the what the rotary encoder controls
(eg, volume control one minute, then over to an EQ etc)


Cheers 'n' Beers :sam:


R

 

[mcs]

[quote author="Rochey"]I agree with you in some respects, although without giving too much away (i'm not allowed to) -- you'd be suprised how many of the uber-high-end console manufacturers use these parts for their analogue desks.[/quote]
The main goal of most manufacturers is to save money - not to make the best possible circuits...

The PGA2311 is a -5V to +5v part, whilst the PGA2310 is a -15v to +15v part. Plus and Minus 15v should be more than adequate headroom - no?

Yes, but it is not a pot - read my post again... It is a volume control, but what is needed is a pot. The three terminals of the pot have to be available for most of these applications, and no op amps are needed or wanted. So chips made by others (like Maxim/Dallas and probably others) are more suitable - does TI make any? But these chips then have the problem that most of them work in the 0-5V range only...

if it's just a pot that you want (not a rotary encoder) then some of the MSP430 devices have an onboard 12bit ADC -- allowing you to sample the voltage through the pot, and having a lookup table on teh device... peice of the preverbial p*ss.

It's not just a pot - the programmability is needed also. Which is why we are talking about displays, up/down buttons, RS-232 and MIDI...

Again, I know I sound a bit too pushy on TI's parts, but all I can say in my defence is that it's really hard working with these parts day in day out, and not having great ideas of where they would fit in audio projects! :wink:

I don't mind TI parts - I use some of them myself. But if other manufacturers have better parts I use those...

Best regards,

Mikkel C. Simonsen

 

[mcs]

[quote author="Mbira"]Glad to see this discussion back! Rochey welcome! So the rotary encoders need to have some sort of LED to see where you are right? I thought I saw some rotary encoders somewhere that act more (physically) like pots where they remember their setting when shut off, etc. Does TI or anywhere else have that kind of thing?[/quote]
You are talking about standard rotary encoders vs. absolute encoders. The standard encoders are up/down only and the microcontroller has to keep track of where the pot "is". Absolute encoders output a unique pattern for each position, so you would just need to decode those patterns (could be done using just a parallel EPROM). But the absolute encoders have the disadvantage that they are expensive. Have a look at the Bourns website. They make both types of encoders. You can find both types of encoders in the RS and Farnell catalogues as far as I remember.

Best regards,

Mikkel C. Simonsen

 

[Rochey]

[quote author="mcs"]
The main goal of most manufacturers is to save money - not to make the best possible circuits...[/quote]

I'm not talking about sub $50K consoles here :grin:

[quote author="mcs"]
Yes, but it is not a pot - read my post again... It is a volume control, but what is needed is a pot. The three terminals of the pot have to be available for most of these applications, and no op amps are needed or wanted. So chips made by others (like Maxim/Dallas and probably others) are more suitable - does TI make any? But these chips then have the problem that most of them work in the 0-5V range only...[/quote]

Ahhh, I gotcha... Apologies... I skim read (which works 90% of the time :wink: ) As far as I know, TI doesn't make digital pots :sad:

[quote author="mcs"]
It's not just a pot - the programmability is needed also. Which is why we are talking about displays, up/down buttons, RS-232 and MIDI...[/quote]
In which case, PIC, MSP430, 8051 etc will all do that. The point i was trying to get accross was that you don't necassarily need an OS to take care of all that.

[quote author="mcs"]I don't mind TI parts - I use some of them myself. But if other manufacturers have better parts I use those...[/quote]

I couldn't agree more. TI, like all of the semiconductor manufacturers has it's strengths and it's weaknesses. (E.g - TI doesn't make great analogue switches for audio, whilst others may)

Cheers 'n Beers,

R

 

[mcs]

[quote author="rafafredd"]I just want a very simple circuit.

I like the idea of two pushbuttons and a three digit display for 0-256.[/quote]
But you wouldn't use 0-255 would you? If you want a rev. log. pot you would probably use 20-40 of the 256 positions. Mr. Bauman please speak up! :grin:

My plan is to make the taper and the display programmable also, so you can display "10" when the pot is at position 27 for instance. That way you can make the display show steps or dBs for instance.

I will make a circuit later today or tomorrow - I'll be back...

Best regards,

Mikkel C. Simonsen

 

[TomWaterman]

This is very interesting - would it be easy to make the relay switching do more complicated attenuators, like T and H pads etc??? I guess it would require quite a lot of relays??

Kev, are you making a midi CC monitor control unit by any chance???
A little remote and a main box with midi switching?? I seem to remember something about this being an idea back at TT.

Also with the digital control attenuators, would it be easy to include memory to have say three or four level settings, I.E speaker A, speaker B ala Cranesong's Avocet?

I would love to use something like this...

Cheers Tom

 

[Bauman]

OOPS!
Mikkel, sorry for all this delay.
No internet at home, at a game lan right now...
I´ll post some photos, sure.

Mikkel has done a amazing job on his digital pot. I think I don´t need anything else. Works just great.

I´ll post some results this next week.

cheers!
Fabio

 

[rafafredd]

[quote author="mcs"][But you wouldn't use 0-255 would you? If you want a rev. log. pot you would probably use 20-40 of the 256 positions. Mr. Bauman please speak up! :grin:
Mikkel C. Simonsen[/quote]

Off coarse I don´t need 256 steps. That´s just because I remenber you saying that it would work with 256 steps with just 8 relays.

But for me, if I had 24 steps that would be enought. 40 is great!

 

[Mbira]

Mikkel,
The digital stuff feels out of my league at this point-OS's, programming relays, whoa! Have you considdered selling a little kit for one pot? I'm sure that if I had one in front of me that worked, I'd be able to understand how it works and recreate my own much easier...If you would be interested in selling one, PM me. I too am looking for a reverse log for an API type circuit.
Thanks!
Joel

 

[mcs]

[quote author="rafafredd"]Off coarse I don´t need 256 steps. That´s just because I remenber you saying that it would work with 256 steps with just 8 relays.

But for me, if I had 24 steps that would be enought. 40 is great![/quote]
You do have 256 steps available - but they are linear steps. So when you want a log or rev. log taper you only use some of the 256 steps. You would probably select the steps so you get an equal amount of dBs for each step.

Best regards,

Mikkel C. Simonsen

 

[kruz]

whoa!!!!!
been away for few days..and...
well..
cool... :green:
actually..
some days ago, me and SmashTV had a great "brainstorming" chat about developping a project for the lab with kits and pcbs avaiable..
i'm taking care about designing the user needs (knobs placements, leds, VU's, buttons, etc..) and SmashTV will do the pcb layout.
The idea is to have a daughter board with buttons, pots, displays, leds on-board, and a second main pcb that could be a simple design (no OS and probably no midi) or a MIOS based version. the user will choose wich one he needs.

I had the same idea as mcs to have an actual readout of the involved dB's, instead of just a steps value.

Maybe, we could team toegheter...
i'll show you all my idea of user layout soon, i'm still working on it.
keep the comments coming!!!
excellent :thumb: :thumb:

 

[mcs]

Here's an updated control schematic:

http://stiftsbogtrykkeriet.dk/~mcs/RelPotC2.gif

Is uses either an encoder or up/down buttons. It includes a two digit display and an EEPROM for storing the "taper". Both the taper and the numbers to be displayed can be stored.

Best regards,

Mikkel C. Simonsen

 

[mcs]

And here's a PCB layout for the above circuit.

If it looks like what some of you want, I'll make a prototype...

Best regards,

Mikkel C. Simonsen

 

[rafafredd]

Looks great, but does it have to be a double sided board?

for DIY a single sided would be better, I think.

So how does the software thing works?

Great work, Mikkel !!!

 

[PRR]

> paranoid about revealing my employer --- ...i'm too biased etc.

In your industry, you must tout your employer's products, but you better not knock the competition because tomorrow you could be working for them. Either hired-away, or corporate merger. There are a few engineers who have worked for about every company in their field, some several times. And buy-outs and product-line transfers.... you never know who you will be working for next week.

Anyway, it is good to have a company man here. Maybe other companies will get jealous and let their folks hang here on company time. A little of that would be a great thing. (Too much of that could be annoying; I doubt that would happen.)

And there are so many chips on the market that it is impossible to keep track of them. Someone like you who sees all the New Product sheets and has a clue about Good Audio can alert us to new toys we have missed. (And I bet you have an eye on products from other companies, and will mention their products when it fits our needs, and doesn't hurt TI's sales.)

I am downloading the 2310 datasheet now. Historically, digital switches had resistance/voltage variations that were no issue when multiplexing into a hi-Z input like an ADC, but meant high distortion when used in the low-Z gain network of a mike preamp. Or you have the Dallas digi-pots, including some dB-taper versions, but many of them low-voltage or uni-polar. A few years back, after surveying what was out there, I liked either the old costly DG switches or the cheap good old 4051 CMOS switch (plus a mess of resistors). Wasn't really happy with either and the idea faded.

And I realize your job is Sales and the fun part of that is New Products, but some of us audiofools are very conservative. Yes, some jump on a new part before the paint is dry; and others think parts from the 1930s and 1940s are best. The key to pushing products to the audio world (not your biggest market!) is to make it "sound smooth", characterize it in ways that audio designers think, and then just keep it in production and pushing samples. Sometimes it takes a few years (or a complementary product) for a part to "click" in audio.

Ah, I've seen the 2310 before. Good chip. Things I question:

Slew rate is lower than I like. The odd bump in the THD/Freq curves may be slew-caused (but I also smell a filter on the THD meter, why does THD drop at 20KHz?). Low slew is not a problem if levels are low, but then I need a post-amp to give Pro Audio levels.

But not too much post-amp, because the output noise, while low (~10uV), isn't very-low. This suggests fussy level setting. Not a big problem coming -off- a recording, but levels and dynamic range can be problems when going from a live mike -to- a recording.

So system level structure needs careful design.

Noise is not specified as a function of gain. I can't tell from the sheet where to start a mike-amp design with this part.

THD will "degrade" if source impedance is over 600 ohms. It isn't clear how much, or even why. That means a buffer in front in many situations.

This is certainly not a drop-in replacement for those pesky inverse-Log pots we use to set gain in mike preamps. It is easy to re-think the topology and it may work very well as a voltage-attenuator to replace the feedback network. While its rated noise is too high for mike-amp, it has real gain (31dB), so its output might be attenuated 30:1 and get unity gain noise around 0.3uV, an excellent performance. (Darn few gain-pot preamps will hit 0.3uV at unity gain; 1uV is more typical for a 5K rev-audio pot at low gain.)

And of course any digital-input part (including the relay boards above) needs a User Interface: a way to turn finger-tips into pulses, and a way to indicate the device setting. For all their other flaws, pots are nice because the user interface is right there.

The 2310 makes a lot of sense when a designer and breadboarder have time to work with it and understand how it best fits their goals. Also when the system already has a CPU, even a minimal brain, to generate those serial codes. For the type of one-off one-channel preamps most folks here are interested in, it may not be worthwhile.

 

[mcs]

[quote author="rafafredd"]Looks great, but does it have to be a double sided board?

for DIY a single sided would be better, I think.[/quote]
No, it doesn't have to be double sided. That just makes it easier to make a small board. I tried to make it as small as possible (it's 55*55mm).

So how does the software thing works?

You program the software thing into the chip and it works (hopefully) :wink:

The 8-pin chip on the board is an EEPROM where the taper/display settings are stored. You can either pre-program a taper into the chip, or you can store one yourself using the buttons.

Best regards,

Mikkel C. Simonsen