Damn, maybe I boobed on the box size then. if so, it'd be the resonance peak rather than the coil inductance causing the grief. I seem to recall that can be flattened with a series RLC. I should have the equations somewhere.
Perhaps a good idea to leave the woofer in the box when I check it with LIMP. I used that to get the T-S parameters, and the resonance "should have been" below the HPF cutoff, but I didn't check it in the box before hooking up the amps 'cos it was "inconvenient".
Probably not so "inconvenient" as having to build another amp, though.
Well, if that's it I have at least learned a lesson today.
Last edited:
Series Notch Filter - Speaker Impedance Correction at Fs (Resonant Frequency)
"This type of filter is generally used on midranges/tweeters. It can however, be used on woofers, but the component values required are large and would be very expensive."
B******s.
Cannot do expensive.
Last edited:
Dear Omega,
It is a bit hard to oversee the situation without being able to make measurements on the amplifier with and without load. This way you never really know what is happening.
If you suspect that there are oscillations, then to start you can stick with the original Zobel, but add a 10Ohm 3 watt resistor with a 5uH (air) inductor in series with the output. (this is the recommended value from the datasheet)
With kind regards,
Bas
It is a bit hard to oversee the situation without being able to make measurements on the amplifier with and without load. This way you never really know what is happening.
If you suspect that there are oscillations, then to start you can stick with the original Zobel, but add a 10Ohm 3 watt resistor with a 5uH (air) inductor in series with the output. (this is the recommended value from the datasheet)
With kind regards,
Bas
Dear Infinia,
Please for a moment stop and don't consider what a speaker is and what their zobels are used for. different concepts! BTW higher woofer series Le is not a problem for amps stability. It affects the amps load line (phase angle) and overshoot (back EMF near clipping ie free wheeling protection diodes)
There are sometimes 2 zobels for the amps stability
1) shunt RC - determined by analysis and/or empirically > usually to compensate for some peaking at light loads on the amps internal EF stage. The value of this Cap is to bring in a low value R to damp the peak before unity gain freq ~ 300 KHz - 2 MHz. Note this is above the audio band.
2) Output small series L with damping resistor. This is used to isolate the amplifier from pure C value on the output. shunt caps or high C ribbon cables and such (sometimes not used and values can range around 1-10 uH ) Also prevents induced RF from entering the amplifier.
That is because folks are not using the the NS recommended RC values. Again youre mixing up concepts and speaker zobels here. Stick with RC zobels in the data sheet <full stop> Get creative on some speaker zobel designs LOL
Ps. Before I forgot. I took your advice and changed ceramic SMD's for film 1206 SMD caps in a DC servo. I tried the tapping test witht the scope. In both cases there was a case of a micro-phony effect with tapping. However when I change the capacitor for a film through-hole part this effect was reduced big time. So I guess I should not use SMD caps anymore in those critical applications. But strange enough even in very expensive "high end" equipment I have seen those ceramic capacitors in PLL loops
Interesting about the SMD film caps I wouldn't expect them to be any where near microphonic as X7R ceramics.
PLLs are my expertise. Depends on loop filter values and many other things. There are good ceramics too like NP0, but they can be very expensive at larger values. NP0 Values up to 0.01 uF and past are rare. If you want to talk about something specific re PLLs, I can tell you some stories too. High gain and/or high impedance sensitive signal lines are to avoided using X7R ceramics.
Okay, I'll check the speaker in the box with LIMP, to make sure the reactive components are as far out of the used used bandwidth as possible. Rebuild a bigger box if necessary.
Then I'll add the parallel RL in series with the amp output, hopefully that'll make everything happy.
Thanks for the advice, you're very kind gentlemen.
Then I'll add the parallel RL in series with the amp output, hopefully that'll make everything happy.
Thanks for the advice, you're very kind gentlemen.
I am there with you, and often adding nr2. to complete the Zobel garantee a stable amplifier with different loads. Many people omit the output inductor/resistor, since it is often specified as an "option" but better be safe then sorry.
Funny you mention the protection diodes from output to voltage rails. I am there with you as well, but I never seen them in the gainclone's or chipamp datasheets. I always add them though
I don't know enough about capacitor materials and requirements, but this tapping test was an eye opener. Seeing those ceramics almost everywhere, made me think that those days they are good enough. It was a wrong "Assumption" from my part. I was sad to see that those SMD film caps did perform so relative poor as well. Make me believe that the SMD technique isn't there yet for quality caps. When I used a simple Wima MKS2 cap, same value but through hole the microphonics where almost no there.
Might make me want to try other SMD caps as well (I can;t generalize all SMD caps because of one experiment with one brand). But I guess it will means that the caps for DC servo's must be through hole then,which is a pitty
Thanks for your insights though.
With kind regards,
Bas
Funny you mention the protection diodes from output to voltage rails. I am there with you as well, but I never seen them in the gainclone's or chipamp datasheets. I always add them though
They appear in TDA2030 datasheet, which is very old. I assumed they had been omitted from later designs due to such protection being provided internally.
Bad assumption, then.
They appear in TDA2030 datasheet, which is very old. I assumed they had been omitted from later designs due to such protection being provided internally.
Bad assumption, then.
Remember application schematics are designed to sell Parts and ideas. The Apps Engineers get in trouble showing big parts counts. Usually reps and distributors say "well your competition only needs X parts how come yours show more".
The Marketing Dept that could not get the Engineers to remove parts somehow made them "optional".
Last edited:
Dear Infinia,
The SMD film-caps I tried weren't Wima's but those:
"http://nl.rs-online.com/web/search/searchBrowseAction.html?method=getProduct&R=0134273" The I was curious right away and those where available in a day. The Wima's are harder to get (fast) here in the Netherlands. Do you think there are SMD film caps out there which can perform as good as the through hole version?
And in how far is taping on a circuit a decent and well argued fact about the quality of a capacitor?
I think they will just tell me "the solution is, don't tap on your circuit" With kind regards,
Bas
Last edited:
Depending on your voltage, I always use simple 4003's or higher rating which do the job. They only need to clamp. a 5408 is kind of overkill
but won't harm either With kind regards
Bas
And in how far is taping on a circuit a decent and well argued fact about the quality of a capacitor?
And keep your amps in a quiet or no vibration zone. "Turn that @#^&& loud music down"
Oh well
Depending on your voltage, I always use simple 4003's or higher rating which do the job. They only need to clamp. a 5408 is kind of overkill
Thanks, I'll check my parts box for 4007's first, then, but I definitely have spare 5408's kicking around. And overkill is good.
Cheers!
But still I wonder if there are good SMD cap's out there, or that I should stick with through hole. The Wima MKS2's really are magnificent in this matter. With best regards,
Bas
With best regards,
Bas
I would design the PCB to take either parts untill you can find a good one.
Every part has a noise floor... with or with out ambient vibrations.
Thanks. I was just wondering you had any recommendation for a good SMD capacitors you have experience with. I just like SMD's since it can make the PCB's smaller.
With kind regards,
Bas
But your PCBs are huge with thru hole anyway
How about telling Wima to give you a sample box of their goodies.
SMD film caps are bigger than their leaded cousins, only dielectric improvements can get them smaller.
Last edited:
But your PCBs are huge with thru hole anyway
How about telling Wima to give you a sample box of their goodies.
SMD film caps are bigger than their leaded cousins, only dielectric improvements can get them smaller.
I can always try to ask Wima and see
The PCB was intentionally designed for use with 1206 size ceramics (my bad), so it was small. I guess for now I stick with the through hole variant, since this performance was really good and stable. And in the meanwhile I experiment with more SMD types.
The ceramics do a good job in PSU decoupling though, and have the benefit with chipamps, you can place them very close to the chips supply pins.
With kind regards,
Bas
I can always try to ask Wima and see
I guess for now I stick with the through hole variant, since this performance was really good and stable. And in the meanwhile I experiment with more SMD types.
The ceramics do a good job in PSU decoupling though, and have the benefit with chipamps, you can place them very close to the chips supply pins.
With kind regards,
Bas
Roger that! I love ceramic capacitors just in the right places
factoid SMD caps self resonant freq is almost cut by 30% if installed sideways ( square body )
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.