|
Post by tupisac on Apr 14, 2017 20:17:58 GMT
Finally Kameleon is soldered and working pretty well. No, actually it's working great The road was little bumpier than with O2, but that was expected. When you move from cozy ready-made kit to just bare board and few pages of instructions things might get a little harder. Most intimidating thing was a filter board. I laughed hard when I saw the caps designated as C4L & C4R: Yep... No way without nice, sharp tweezers and soldering tip. Also, Frans technique of inserting one side to molten solder proved essential. Flux in syringe with needle was also very helpful. Overall - very nice experience and definitely worth the effort.
|
|
juke
very active
Posts: 396
|
Post by juke on Apr 14, 2017 22:13:38 GMT
That's one huge match
|
|
|
Post by tupisac on Apr 14, 2017 23:27:30 GMT
Wanna see my huge jack?
|
|
|
Post by tupisac on Apr 16, 2017 20:04:21 GMT
In all fairness and jokes aside, those tiny capacitors are only elements this small. All the others are at least twice as big. Strangely, they are all designated as 1206 - even the caps. I assumed it was size, but somehow it isn't. Or there was a mistake on Conrad's side. Now it's time to get serious with the enclosure. Here is the latest design, keeping with the theme from Polaris: But there is a slight problem: The screws and distances from Polaris. By the strange measurements I assume those are in freedom units. Slightly smaller diameter than M3, but with less dense thread. And those distances - I've only seen ones that are hexagonal. Anyone know where can I get round distances in EU? Preferably with M3 thread.
|
|
solderdude
Administrator
measureutternutter
Posts: 4,882
|
Post by solderdude on Apr 16, 2017 21:04:35 GMT
that SMD part looks like it is 0603 sized. Most likely the value is correct (10nF) but may just have been the wrong size.
You can contact Jeremy and ask him to send you a set of screws/nuts/stand-offs in a bubble envelope. Then it will match the Polaris. Will probably be UNC size (bolts)
|
|
|
Post by tupisac on Apr 17, 2017 19:35:04 GMT
The SMD part has 1206 written on a bag and in description. Seems like Conrad's fault.
I've found round stand-offs in black with m3 thread. RC and drone shops offer those.
There is one thing that bugs me a bit with Kameleon. The filtered audio is much louder that bypassed. Is it possible to assemble the filter with gain=1?
|
|
solderdude
Administrator
measureutternutter
Posts: 4,882
|
Post by solderdude on Apr 17, 2017 21:17:50 GMT
The gain in filtered condition should be around '1'. There should be no difference in perceived loudness, only in tonal balance. This is determined by R1 on the filterboard, the inherent gain of the filter and the 470 Ohm resistors.
Do you know what values are used on the board for R13 and R23 on the board and R1L and R1R on the filter PCB. I assume the HD600 module ?
|
|
|
Post by tupisac on Apr 17, 2017 22:31:39 GMT
Yup, HD 600.
R1R, R1L: 470 Ohm R13, R23: 470 Ohm
I was initially comparing it on music. Those need like 1,5 hour turn on pot to match. When I checked with the 8, 9, 10 and 12 kHz signals the difference is much smaller. But it is audible. Filtered is slightly louder and a bit "wider". Shouldn't those frequencies be a tad quieter when filtered?
EDIT: Confirmed with the spares from the bag of R13 and R23: 468-469 Ohm. Same with R1R and R1L. Also those have 4700 written on them.
|
|
solderdude
Administrator
measureutternutter
Posts: 4,882
|
Post by solderdude on Apr 18, 2017 6:47:45 GMT
Acc. to the simulations the loudness should be about the same. Maybe 1dB too loud which is audible on direct comparison. Because there are also psychoacoustic things going on in a filter it is hard to determine equal loudness so you could try 560 Ohm for R1R and R1L on the filterboard. Don't have a HD600 nor a filter for it here so cannot try it.
|
|
solderdude
Administrator
measureutternutter
Posts: 4,882
|
Post by solderdude on Apr 18, 2017 10:51:49 GMT
I can send you some 0805 resistor values so you can experiment with the values of R1 and see which value 'fits' best.
|
|
|
Post by tupisac on Apr 18, 2017 20:34:21 GMT
Thanks, I'm good. I've got the electronic shop nearby and they have SMD caps and resistors. Just tell me what values should I get. Senses lie, so I went more "scientific": Here are the measurements: Signal
| Filtered (V AC)
| Unfiltered (V AC)
| 100 Hz
| 0,118 | 0,056 | 500 Hz
| 0,1 | 0,055 | 1 kHz
| 0,097 | 0,054 | 2 kHz
| 0,079 | 0,05 | 5 kHz
| 0,012 | 0,009 |
According to the FR charts from filter description those frequencies shouldn't be affected this much by filtering. What are those other psycho-acoustic things going on with the filter?
|
|
solderdude
Administrator
measureutternutter
Posts: 4,882
|
Post by solderdude on Apr 18, 2017 21:55:22 GMT
Weird,
Are R10,R11,R20 and R21 all 10k ?
These resistors lower the input voltage by half. The filter itself will have 4x gain (for usage with old Kameleon amps and 2x for Kameleon2) As the input voltage is halved you have 2x gain left (+6dB) R1 with R13 half the output voltage yet again so you are left with 0dB gain (=1x)
So it is strange that you still have 1.8x gain.
When R10 and R20 are 1k for instance and R11 and R21 are 10k one would expect 0.9 x 4 x 0.5 = 1.8x gain. Perhaps check those resistor values.
With the HD600 filter the filtering is actually quite subtle except for the (sub-bass) With some filters parts between 300Hz and 6kHz for instance may have considerable changes. The ear reacts different to gains in different frequenvy bands when it comes to perceived loudness. With some filters it is thus not easy to calculate the level difference.
|
|
|
Post by tupisac on Apr 18, 2017 22:03:58 GMT
All 10k. How to measure them in circuit?
|
|
|
Post by tupisac on Apr 18, 2017 22:24:17 GMT
I specifically chosen the frequencies according to this chart: Those actually should be lower or equal when filtered. All the components are according to revised BOM. There is a big possibility of me screwing something up, since it was my second board soldered in my life. But not with the values. More likely is that I might have destroyed some components with prolonged heating (temp 330 C). How to probe the vital components and filter elements in-circuit for that?
|
|
solderdude
Administrator
measureutternutter
Posts: 4,882
|
Post by solderdude on Apr 19, 2017 5:00:04 GMT
Strange.. Highly unlikely the parts on both channels became defective in the same way.
The only 'suspect' part may be the small sized cap which MAY not be 10nF. When it would have a much larger value this could well explain the difference in gain and yield a wrong correction as well. You could simply remove that part and see if the gain alters. Also you could check the resistance values on the filterboard as these can be read.
I have 10nF SMD caps which I can slip in an envelope for you.
Also noticed the filter BOM list is not complete. It misses the 3kHz dip circuit. Will try to correct this soon and upload the filtermodules.
Looks like C6, R7 and L1 are not in the BOM list.
|
|