24 dB/Octave 2/3-Way Linkwitz DC on output after power cut

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
oh no!

i was wrong, it was not that small peice of solder, that was the problem.... it did the same thing yesterday:(

so i now im trying to find out what is wrong. using a scope and a signalgenerator.

one of the channels works fine, the other one do that "thing"

to fint the fault i have connected both channels input to the signal generator. i tryed to measure at the outputs, on the channel that i think is okay. it looks fine when i feed it with a sine wave.

but when i try to use a triangle wave!:bigeyes: at high freqency (8000Hz) it looks nothing like a triangle wave more like a sine wave with a lage deformity.. at low frequensy it very hard to tell if it is a sine, triangle or square wave.. they all look almost the same :confused: that can't be the intention?? when i try a square wave at 8K Hz

Help plz:dead:
 
I have built Rod Elliot's LR4 using his board without any problems. Indeed it sounded quite well. The turn-off noise was so small it did not bother me at all. In any case, I always turn it on before the power amp and turn it off after the power amp. This is very well documented by Rod on his protected site.

I used Vishay / BC Component precision film (polypropylene) caps. In any case, I still found some "veil" in the music.

I then did some upgrade on Rod's board. This was to use better local filtering by using a 50uF 35v Panasonic FC plus a 0.01uF ceramic NPO on each of the positive AND negative rail for each half of the OPA2134. It was a hard job because Rod's board is pretty tight but the job is not impossible. Let me tell you that this improved the sound greatly. This was using Rod's preamp power supply.

In later events I tried adding the baffle step compensation and other filters therefore I ditched Rod's board but used his schematic for the LR4. But what I found was that soldering the ICs on veroboards can easily overheat the ICs (believe me, sorry I don't have more time to write more on this) and cause damages. So I would highly recommend any beginners to use Rod's board for a start because you are almost guaranteed a success. I later found baffle step compensation for my speakers was unnecessary on the active step and it all went back to Rod's circuit.

That is not to say you can't do further improvements over Rod's board. At the output where it allows gain adjustment, I changed it (Rod also gives some instructions on how to do so on his document). But I changed it in a different way. Rod uses a pot and have a gain stage. Since I have the low pass and high pass and I know exactly which one needs gain or attenuation so I only needed to install a pot to attenuate either the LP or HP. In this way, all OPAMPS are used as unity gain buffers and I actually save one OPAMP on the signal path.

Another improvement I made was this: no drivers are flat so I take advantage of the natural rolloff of the driver (6dB). So in order to achieve an acoustic 24dB order I need only 18dB. 12dB is achieved using one OPAMP. The other 6db is simply a capacitor worked out to be a value to form a 6dB roll off with the input impedance of the power amp. So this is what I have done: on the HP (which is far more important) the first OPAMP is the buffer amp. Second is for the 12db. I have only 2 opamps in the signal path but I have achieved 24dB rolloff with minimum number of capacitors as well. I saved the output cap of the LR4 and the input cap of the power amp. The sound is of course much better.

I also upgraded from Rod's basic power supply to one of my own design / implementation of the LM337 data sheet. let me say I am very happy with the sound and this is high-end.

I think you can get away using a standard active 24dB LR XO on low frequencies (not higher than 500Hz, e.g.) but I would not recommend using it for XOs between tweeters and the midranges because it will never work well! Do it if you have very advanced skills and measurement equipments (and I don't) otherwise, GO PASSIVE! i.e. passive between the tweeter and the midrange, and active between the midrange / base and woofer (and XO at lower point) - this is the best of both worlds.

Regards,
Bill
 
lets see what the scope tels us

I hoked a scope up the output, in this case the treble output.

this is a 8Khz sine wave:
 

Attachments

  • 8khz sine.jpg
    8khz sine.jpg
    58.3 KB · Views: 222
I had the same doubt about my active 24dB Linkwitz when measuring with a scope. To confirm the measurements I found a circuit simulation program on the net and emulated the xover on the PC – :bigeyes: to my surprise the measurements was confirmed.

If you look in the DIY literature you will find that the drawback of a 24 dB xover is poor puls response, I think that is what the scope is showing. Fortunately my xover sound a lot better than I “looks”.
 
oh i was not aware that. strange! it looks so bad, but it sound good? i don't understand it is posible, anyone care to explain?

another thing i do not understand is when i turn the frequensy up to a few hundred kilo Hz, then thing starts to look fine agin? how can what be?

tnx
 
Hi Bill,

I haven't been following this thread for a while, so sorry for a question this late. I'm working on the ESP 2-way LR XO. I follow Project 09 exactly. There are 5 opamps per channel - 1 as input buffer, 2 for HP and 2 for LP. As I understand, each pair is a cascade of back to back butterworths. What you're saying is to use only the first butterworth and a cap of appropriate value to replace the second to achieve the same 24dB roll off. Right?

Cheers,

KK
 
These waveforms are not right. Sine waves should never get distorted when they pass through any crossover. Forget all these fairy tales about "transient degradation". The most probable cause for that distortion is an op-amp in some filter stage clipping or driving too low an impedance. It may be also due to op-amp instability.
 
kkchunghk said:
Hi Bill,
What you're saying is to use only the first butterworth and a cap of appropriate value to replace the second to achieve the same 24dB roll off. Right?


I think Bill meant: 18dB roll off from the active filter + 6dB from the driver = 24dB.

This is in fact a good point to remember: the total response is the sum of responses from the active filter plus the driver responses – and other forms of unlinearity in the chain. Therefore, if a standard active xo is used, the drivers have to be inside their working range and the response fairly flat. E.g. the xo point should be well away from any brake-up frequencies. Otherwise the filter has to be customised – as Bill does.


To EVA

I mentioned earlier in this thread that my 24dB active filter distorts pulses and it do. But the sine waves are not distorted, and that’s unquestionably the most important issue. IMO transient response also has an impact on sound reproduction, but surprisingly modest.
 
kkchunghk said:
Hi Opp,

Let me try again.

24dB = 6dB of the driver + 12 dB from one opamp + 6 dB from a cap

This is how I understand it for I'm short of board space and desperate to get rid of an opamp. That's why I was focusing on Bill's mentioning of one opamp.

Cheers,
K K


Oh……… sorry, that could be done.

But……. Using the speaker roll off and the cap/amp.-impedance roll off, might produce some kind unwanted effect. The result would be a little unpredictable, especially: how is the driver performing in the roll off region, and how much will the roll off change over time? The best advise is to build a test set-up and listen to it for a couple of days. :violin:
 
opp said:
... my active 24dB Linkwitz .....
..... drawback of a 24 dB xover is poor puls response ....

It is not difficult to change your active filter to any other type.
It is just another formula to set values of resistor and caps ( R and C ).

Butterworth is a good allround filter.

Linkwitz-Riley filter characteristic is know for good phase response.

Bessel filter is used for improved pulse response.

Chebyshev filter, I do not know what is good for.

Bessel and Chebyshev is not so often used. Only for special applications.


If a Bessel 24dB filter has any good pulse response( compared to 12dB), I do not know.
But I guess it is a bit better than Linkwitz-Riley, which is optimised for good phase behavior.
 
Some info on Bessel filter, in comparison with Butterworth and Linkwitz-Riley.
What Is a Bessel Crossover?

The Bessel filter was not originally designed for use in a crossover, and requires minor modification to make it work properly. The purpose of the Bessel filter is to achieve approximately linear phase, linear phase being equivalent to a time delay. This is the best phase response from an audible standpoint, assuming you don't want to correct an existing phase shift.

Bessels are historically low-pass or all-pass. A crossover however requires a separate high-pass, and this needs to be derived from the low-pass. There are different ways to derive a high-pass from a low-pass, but here we discuss a natural and traditional one that maximizes the cutoff slope in the high-pass. Deriving this high-pass Bessel, we find that it no longer has linear phase. Other derivations of the high-pass can improve the combined phase response, but with tradeoffs.

Two other issues that are closely related to each other are the attenuation at the design frequency, and the summed response. The traditional Bessel design is not ideal here. We can easily change this by shifting the low or high-pass up or down in frequency. This way, we can adjust the low-pass vs. high-pass response overlap, and at the same time achieve a phase difference between the low-pass and high-pass that is nearly constant over all frequencies. In the fourth order case this is 360 degrees, or essentially in-phase. In fact, the second and fourth order cases are comparable to a Linkwitz-Riley with slightly more rounded cutoff!
....

Summary.
It is seen that a Bessel crossover designed as described above is not radically different from other common types, particularly compared to the Linkwitz-Riley. It does not maintain linear phase response at higher frequencies, but has the most linear phase of the three discussed, along with fairly good magnitude flatness and minimal lobing for the even orders. It is one good choice when the drivers used have a wide enough range to support the wider crossover region,
and when good transient behaviour is desired.

From a Rane note:
A Bessel Filter Crossover, and Its Relation to Others

n147fg11.gif

Fourth-Order Group Delays
 
lineup said:


It is not difficult to change your active filter to any other type.


Unfortunately mine is :clown:, it was only a test muck up, but has bee in service for a year now– otherwise you are right.


I have been thinking about the term “pulse response”, maybe there are two different types:
1) Electronically, that what we se on the oscilloscope. ms range.
2) Musically, e.g. a guitar being strummed. 100ms range.

This might explain why (Electronically) impulse isn’t that important.
 

Attachments

  • billede 004.jpg
    billede 004.jpg
    38 KB · Views: 160
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.