Hi guys,
Yes, I know there's a sticky about this topic, but no-one i've asked so far seems to be able to give a straight answer for a starting point due to the unusually low input impedance.
I'm populating some DIY stepped attentuators for going between my DSP and these boards, precision volume control is not a concern, think of it more as gain control so I can stamp out any noise floor from the DSP and set a safe max volume.
Can anyone suggest what value of pot I should be aiming for. From reading the sticky I'm thinking that 5k would be good, but I wanted to check with you smart fellas first.
No, I won't be adding input buffers or anything else, just passive volume control.
Thanks in advance!
Yes, I know there's a sticky about this topic, but no-one i've asked so far seems to be able to give a straight answer for a starting point due to the unusually low input impedance.
I'm populating some DIY stepped attentuators for going between my DSP and these boards, precision volume control is not a concern, think of it more as gain control so I can stamp out any noise floor from the DSP and set a safe max volume.
Can anyone suggest what value of pot I should be aiming for. From reading the sticky I'm thinking that 5k would be good, but I wanted to check with you smart fellas first.
No, I won't be adding input buffers or anything else, just passive volume control.
Thanks in advance!
Frankly, I'd use a 10k pot and an OPAMP buffer. Generally, the pot should be 1/10 of the input it's driving - AKA a 180R pot for the 1k8 input - obviously quite a silly idea because only a headphone amp could drive it.
How come the example in the sticky post ended up with a 5-10k pot on a 3.2k input impedance amp?
It has to do with HF rolloff but low values should be ok unless the input capacitance of the amplifier is high.
Are you sure you've got your maths the right way round, he should have needed a 320r pot by your method.
kodabmx Math is right, he states what your mystery user "should" have done.
That he can get away (sort of) with something else is completely different.
FWIW IcePower themselves do NOT suggest wiring a passive pot at the input ... they "should" know.
In the datasheet they show a low impedance preamp directly feeding Power Amp input, no potentiometers anywhere.
That he can get away (sort of) with something else is completely different.
FWIW IcePower themselves do NOT suggest wiring a passive pot at the input ... they "should" know.
In the datasheet they show a low impedance preamp directly feeding Power Amp input, no potentiometers anywhere.
Why?
it's not a "mystery user" it's the stickied post at the top of the board. Are you saying a post that's been pinned for over a decade has been fundamentally wrong all along and no-one has brought it up?
The guidelines also show what is assumedly a pot, pic related, circled in red.
I've just run the maths, and with a 2k pot attached, the minimum load the pre will see is 1k, 10 times the output impedance of my pre (motu m4, 100 ohms), and whilst the volume curve won't quite be ideal, it will be very usable, I may swap over to a linear array as opposed to log, as I plan on having it as near to max volume as I can whilst squashing out any noise, so finer control in the top 1/3 is much more important to me than the bottom 2/3.
I don't wish to add an input buffer as I do not want to colour the sound or degrade the signal at all, the primary purpose of this amp is for testing loudspeakers.
Attachments
Wherever he is, he´s not here.
1) there´s TONS of stickied posts, you don´t say which. (you must link it here)
You do not mention the user. (Same thing)
You do not mention then specific post saying what you say they say 🙄
You do not quote the paragraph saying what you say they say
That qualifies as "mystery user" any day of the week.
You are contradicting a sensible recommendation, the burden of the proof is on you.
- Where do I say that?
- Which post?
- In any case, don´t use straw arguments, don´t "are you saying" me, it always ends wrong.
Where is the pot?????
I see a balanced shielded wire connection.
You are confusing output impedance, an "internal" parameter, usually valid in "small signal" conditions, and load impedance, an "external" large signal parameter which will really tax your preamp.
I doubt it can cleanly drive 5V RMS (+16dBu) into a 100 ohm load.
And IF it does, it will be very unhappy.
Even better, an example: a 100 damping factor power amp, rated for an 8 ohm load, will have an internal impedance of 0.08 ohm
Do you expect that amp to safely drive 0.08 ohm? With the full output swing?
Or even 0.8 ohm?
I see you did not "do the Math"
Under your best setting, "1/3" on a linear scale, you are attenuating preamp signal by 3.6 dB
So you are attenuating already microscopic noise by an extra 3.6dB
Paying he price of disrupting Pre to Amp interface big time, Amp will not "see" 100 ohm any more but, say, 10X that, thanks to the added, un needed, not recommended pot.
Think about it.
You are achieving the exact opposite, and messing big time with that interface.
For a "gain" of nothing, in practical terms.
PS: maybe you didn´t notice that for that balanced interconnection you will need a dual pot.
PS2: I suggest you just straight run that nice preamp output into that nice power amp input,using the recommended balanced cable and connectors, it will work fine for your purpose.
Don´t get entangled in "perceived" advantages which later turn up to be handicaps.
And by the way, your Preamp already has a dreaded Buffer at the output 😉
Last edited:
I had an AS100 and used a 10K pot at the input. It worked fine but I was using it below 700Hz for a woofer. The module did need a hefty level to drive it.
IIRC, I just grounded the negative terminal. I also ran it balanced with no pot, just digital volume control and that worked fine too. I liked it a lot, very strong amp.
IIRC, I just grounded the negative terminal. I also ran it balanced with no pot, just digital volume control and that worked fine too. I liked it a lot, very strong amp.
There's ONE stickied post at the top of this board.
Pic related.
Nice essay though bruv, why am I not surprised you watch Jordan Peterson 😂
If it was just a hifi amp I'd have a lot less concerns about volume control, but in my experience changing DSP settings often leads to loud pops that pay zero attention to what the digital volume control is set to, and I don't feel like breaking any more tweeters. But also turning the amp off every time I want to change something on the dsp is impractical and makes A\B comparisons much less effective. So whilst daily listening the amp will be turned up high and digital volume down low, whilst tuning I will do the opposite.
I've sent Motu a tech support request asking what minimum impedance the line outs can happily handle is, if it can handle 1k then I think a 2k pot should give me more than adequate results. If I have to add a line buffer I'll need to add 8 channels worth, and that all adds up very quickly, not only in cost but also enclosure space! 😵💫
Thanks for your insight though, I'll go and do some more reading and try to find out if I'm going to have any problems with HF rolloff.
Pc based is usually fine (but reliant on the software always working as intended), minidsp 2x4hd sometimes popped, especially with fir filters, adau1701 with sigma studio based DSP units I've used have given me all sorts of pops, whistles and hisses on changes. Ironically I think my old Beringer ultradrive was the best unit I've had in this regard.
I've been using DSP units for speaker design iteration for 20 or so years, this is my first time building a proper amp box for it all so want something that ticks all my boxes.
If it was a proper amplifier, you wouldn't be trying to cut corners by omitting a 10$ buffer. What do I know though?
It's not a "proper amplifier", it's a speaker testing rig. "Proper amplifiers" tend not to have 8 channels either 😉
The IcePower modules have a very high input capacity - when you put a resistor in series you loose high frequencies! There is some information about this in a design manual of the amp ... can't find it at the moment.
I had to modify a MiniDSP plate amp for one of my (older) projects cause there was to much hum and noise, you can rather easy replace input resistors and capacity (but it looks nicer when you have parts in the proper size on hand 🤓 )
I had to modify a MiniDSP plate amp for one of my (older) projects cause there was to much hum and noise, you can rather easy replace input resistors and capacity (but it looks nicer when you have parts in the proper size on hand 🤓 )
Oh that looks like a fun bit of soldering 😐
Do you know if this is the case for all of icepowers modules? I've noticed that other models in the asc range that I have, have normal, much higher input impedance, which makes me believe these are more likely remarketed designs originally made for 3rd parties, rather than actual ranges of products designed with compatability in mind.
Not sure about the newer ones, but the original ASC-modules were designed for lower-cost applications (e.g. consumer active subs) and so they got a different feature-set on purpose. That included a lot of the features you'd normally put on a front-end so that it would be simpler and cheaper for speaker manufacturers to implement into end applications.
The "original" meaning of the ASC acronym is actually Amplifier + Supply, Consumer (vs. ASP = "Amplifier + Supply, Professional" for the other original discrete amplifier series).
You could have a look in the datasheet - it says 220pF input capacity. Now you can calculate your f3 frequency and loss at 20kHz dependend of your series resistor.
The input impedance of that module is unreasonably low. It's obvious from the stated input capacitance that the impedance falls with frequency, which means that if you DON'T use a buffer you WILL be coloring the sound.
It's funny how common misconceptions are exactly the opposite of reality. This application virtually demands an input buffer. Two side benefits are that you can standardize the input impedance to something more reasonable (like 10K or 47K), and that your volume control will work much more precisely.