This is unfair and unwarranted.
It is NOT an OpAmp problem, at all, but poor implementation and/or: User Error.
Improperly using that balanced input makes the useful 330pF RF interference capacitor into a 20kHz |attenuator.
And even so ... guess 10 to 20db less than typical user attenuation at that frequency.
No, I simply didnt get to read it before posting the FR performance of the test interface, as that seemed to be requested many times here. Then, instead of leaving the post at that, I blathered on about the circuit. Sorry...I do appreciate all of everyone's effort to right my understanding fo what's happening here!
I initially had the opposite input shorted, then later removed it thinking it wasnt needed, as there's a bleed resistor to reference to ground there anyway. Let's see if the op-amp diff to SE converter measures properly with the (-) input dead short to gnd, signal going into the (+) side.
My mistake, not understanding how to wire a SE output properly to a balanced input. Actually I did understand at first, but then overthought it in grasping at straws.
The NRV10 has balanced outputs and I could have wired up a couple canon connectors for the initial listen; but, I thought, I'll never use the amp that way, only with SE outputs, so why bother.
This loopback measurement does not show this heavy roll off at 20kHz from first post.
IMHO the rane amps do not have such a ultra high deep supa dupa filter at 20kHz.
Was the sampling frequency in both measurements above 96 khz and no digital filter active? deemphasis deactivated et al?
BTW, measure 3times: loopback, left and right from the output inductor with same settings.
Have fun, Toni
Last edited:
Please read #51. (I don't know how we got to 50 posts without this key clue.) It never was "loopback error", and we knew that was unlikely from #1 ("another SS amp" showed no rolloff). But there was a whole other stage in front which did an unexpected thing when connected unexpectedly.
Old wire-benders like me might have thought to hard-reference the unused screw. A transformer input would have gone very wonky. But who uses transformers? Or understands how the RF cap cross-talks the diff-amp above 10kHz?
This time I made sure that ground jumper has continuity by explicitly checking across the pins with my meter. I may very well have gripped the wire jacket with the screw-down jaw of the connector in my previous SE to Bal connection attempt.
Going across all 6 amplifiers, each measures 20-20k -1db at each end of the spectrum. I for one did not "understand how the RF cap cross-talks the diff-amp above 10kHz" until going through this exercise. Now I know and hopefully remember this experience.
I had a listen tonight - same. Confounding things, I suspect the tweeters on the HLS 610s are blown. Imagine a timeline like;
Joe burns his tweeters, but doesnt know it --> gets new amp, doesnt sound right --> measures an erroneous connection's effect he doesnt understand how could possibly be --> ascribes sound to measurement --> posts in DIYAudio about it --> ratifies his measurement equipment --> gets straightened out why this particular balanced input circuit with lopsided connection...-->fixes input connection, all re-measures as OK -->still doesnt sound right -->other Kenwood amp doesnt sound right now, either.
The HLS are hanging from the rafters. If I go up there on a ladder to pull out the connector panel / crossover and measure the tweeters continuity, I just might fall off and really have something to write about. I do have replacement tweeters and I dread taking those back down, but I've had enough circuit sleuthing for one weekend.
Playing around with 100 WPC amplifiers with no input level controls...well, lets just say that screech I heard when the Kenwood was connected as the RCA cable pulled itself out of the RCA to 1/4" adapter in the back of the NRV10 - probably could have done it. I swear I connected the Kenwood back up in the middle of that timeline and it was sounding fine through the HLS; but maybe I only measured it.
Going across all 6 amplifiers, each measures 20-20k -1db at each end of the spectrum. I for one did not "understand how the RF cap cross-talks the diff-amp above 10kHz" until going through this exercise. Now I know and hopefully remember this experience.
I had a listen tonight - same. Confounding things, I suspect the tweeters on the HLS 610s are blown. Imagine a timeline like;
Joe burns his tweeters, but doesnt know it --> gets new amp, doesnt sound right --> measures an erroneous connection's effect he doesnt understand how could possibly be --> ascribes sound to measurement --> posts in DIYAudio about it --> ratifies his measurement equipment --> gets straightened out why this particular balanced input circuit with lopsided connection...-->fixes input connection, all re-measures as OK -->still doesnt sound right -->other Kenwood amp doesnt sound right now, either.
The HLS are hanging from the rafters. If I go up there on a ladder to pull out the connector panel / crossover and measure the tweeters continuity, I just might fall off and really have something to write about. I do have replacement tweeters and I dread taking those back down, but I've had enough circuit sleuthing for one weekend.
Playing around with 100 WPC amplifiers with no input level controls...well, lets just say that screech I heard when the Kenwood was connected as the RCA cable pulled itself out of the RCA to 1/4" adapter in the back of the NRV10 - probably could have done it. I swear I connected the Kenwood back up in the middle of that timeline and it was sounding fine through the HLS; but maybe I only measured it.
It's a pity this didn't solve the bad sound. Let's hope it's the tweeters and not hearing damage caused by the screech.
Imagine? wishful thinking!!!!!
Been there, suffered that, like most here.
Deep Philosophical thought about it: sh*t happens 🙁
I think I'll put some pink noise through these speakers and hold up a measurement mic into a 48th octave analysis - while standing on the floor - much safer approach. I really have to watch the attitude with which I get up on a ladder; exasperation around what's happening isnt a good mode to be up on one of those...
Impressive. I'll keep that value reduction in mind as a possible candidate for change. Probably wont touch C8 though.
Oh, like those Bootstrap capacitors for the HS fet of a class D - I get it. There's something to leave be.
While checking each amplifier with the unused side of the balanced input properly grounded, I did put the two 8 Ohm loads in series and measured the last one again. It was about the same; even though I now know that L3 is not the issue, I did the experiment anyway.
Interesting. The amp has switches so someone can run two amps in BTL for 300W supposedly; perhaps the bootstrap isnt "only one half of the sinewave" in that mode? A mode I'll never use and an output voltage level I'll never take it to where the driver transistor needs a supply greater than B+.
Thanks for doing that. You can see in the board photo the inductor. Now I cant tell inductance from physical size / turns in air, but I think this is element getting ruled out!
I have a Focusrite Scarlet solo, which I havent setup yet for doing audio signal measurements. When I do, the brick wall filter around 20k wont be a concern. My next focus will be getting the harmonic distortion spectrum response and the Focusrite with its much better sampling performance will allow me to see some of those out to 20k. But tomorrow I'll use it to verify my blown tweeters, vs ears / brain.
The designers didnt seem to have any qualms about how many op-amps or caps the audio signal went through. I'm still on the fence, but I'll probably just leave the balanced input intact - even with its idiosyncrasy when driving single ended signals.
The particular Kenwood amp doesnt have one - just the RCA input and speaker outputs. Was planning on drilling a hole in its front panel (I got it for $5 at the goodwill outlet, cover is is a little too bent up to sell; judging by the price they go for, no one would want it anyway so who cares?) and slamming something in there, but never got round to it - and now this.
By acoustic measurement and comparison with a known good speaker, the HLS tweeters are dead. Dammit. The coincidence got me all suspicious about this Rane amp, when it was something else altogether. I guess we can add deaf, to dumb and blind.
I noticed in one set of xovers I've dismantled, they used incandescent lamps (in series) as tweeter protection. Better to light up one of those, than the tweeter VC on accident. I'd like to add something to protect my last tweeter set when I repair these. The Rane has volume controls per channel so it's a bit safer than that Kenwood.