Setting up a ESP active crossover with gainclone

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Yep, found the formula, did a few calcs, although could parallel the caps I was looking at worst case scenario (speaker is rated at 4ohms but not sure of individual drivers - unlikely but 2ohm each? I know absoltulty nothing about passive crossovers at present)

Therefore I would be looking at a 99.5uf cap, and when it comes to non electro's they get quite pricey! (actually its probably not, im just used to trying to find high voltage AC rated examples)

I will most certainly be giving it a go to see what it sounds like, im just stuck at work at the moment and was intersted to pre-empt what might happen.

Regards

Dan
 
Hi,
being British retail, the speaker is less likely to be 4ohms, more likely to be 4 to 8ohms, meaning a 4ohm bass/mid driver and an 8ohm treble driver.

If it has a 4ohm bass/mid driver then you become quite limited on what you can do with the chipamp supplying the bass/mid.
 
Right well I've sorted the 4ohm speaker issue by building a pair of bookshelves with the hivi d6.8 (8 ohm) and sd 1.1a tweeter - should be finished shortly.

Finally finished the 2nd set of amps as well and have begun testing using the jamos.

At certain volume levels or at certain levels using the trim pots a contstant ermm hiss and cripsy crackle (low level ish) comes through the tweeter of the channel being ajusted. It seems to build up as I'm ajusting in one direction and then drop down again to almost nothing once it has been passed.

Can anyone hint me in the direction of what causes this effect/ proposed rectifiction or has anyone else experienced this? Im using a 50k alps log for volume and 20K trimpot as sugested.

Oh and hint for anyone using the ESP active crossover, there is a error on the silkscreening of the board (LP and HP the wrong way around) which is noted in the manual but doesnt jump out if in a rush!

Regards

Dan
 
I have 3 spare so I guess I could. Will have another look into it tonight, I have a feeling it occurs on both channels though, and only one opamp is shared between the two channels (the 1st, the other, 12 are channel specific - hope that make sense)

I did pay attention to ensure minimal heat was applied for shortest duration possible whilst still achieving satisfactory join.

Complete stab in the dark (no idea what im talking about here) but its nothing to do with the output impednce becoming a value which somehow interacts with the amplifiers?
 
Pinkmouse,

Circuit is 100% as per Rods 2way design.

Have tried swapping them over and it still occurs, I can confirm this happens on both channels ( now that I used a SPL meter to get the outputs of both channels matched)

It appears many other people have experienced such hiss but cannot find a definative answer to it.

At the current levels (trimpot positions) as the volume is increased it sounds like one is tuning a radio in (that varying tone). When 12 oclock is reached all his disspears and the whloe system sounds perfect, when around 3 oclock is reached the hiss comes back for the last 1/4 turn. (however all hum in the mid/low amp diminishes)

I also tried the amp with my new tweeters and the same thing happened.

Its driving me mad as I have quite limited time before I go to university!

I breifly tried a 1uf (didnt have any 2.7's to hand) / 1ohm zobel and the problem was still there (if it made a difference it was only very slight).

For the 1st 5 - 10 seconds the problem does not exist, but them builds up to a steady level.

shorting the inputs on the active crossover makes no difference

All help and input much appreciated!!
 
Ahah didnt know they had one! Just waiting to be authorised by an admin.

I know you have said that you havent got a clue on this one..

But this morning I tried the x-over with my surround receveiver and guess what, no hiss!

Did measure an increase from around 0 (using 20 Meg range on meter) to 14Mohms within the first 5-10 seconds on the outputs, which co-incides with the introduction of the hiss on the gainclone.

Therefore both amp and xover are happy on their own as long as they are not brought together! Which kinda sucks!
 
They are Peter Daniel LM3875 Premium GC's.

By low pass I take it we are refering to a Zobel?

Well I tried inserting a 1uf and 1ohm resistor (peter reccomends 2.7uf but I didnt have any to hand)

The problem was still there after this. If i made a difference it was only marginal.

No DC blocking as far as im aware, currently wired as per PD's instructions. - However the 1uf caps on the output of the crossover should take care of this?

I tried a CD with 0 to 20Khz sweep on it and it either made the amp unstable (sounded like clipping (or the CD was skipping - I should test it in another player) but more of a pulses white noise essence to it) - However this has only happened once during general music playback
 
Ex-Moderator
Joined 2002
That's the output Zobel, that shouldn't affect things that much. PD's circuit is very nice, but it does lack a few things that can help stability, as I found out when experimenting with GCs in my old house, just a mile away from the Crystal Palace TV transmitters.

Okay, there's a couple of things we can try, first thing is to dowload the datasheet. If you look at the typical applications schematic, you will see a 220p cap across the inputs. You can add that very easily just by tacking it onto the bottom of the board and it reduces the amp's sensitivity to RF, not so much a problem in the wilds of Canada, but it can be an issue in the EMI crowded UK.

The other thing to try is the resistor and cap to ground shown at the tail of the feedback network. This reduces the DC in the feedback and makes the amp more tolerant to different preamps.

I'm not saying that adding these will solve your problem, fault finding at a distance isn't straightforward at the best of times, I much prefer to have the circuit in front of me with the scope and signal generator fired up and ready to go. ;)
 
Ex-Moderator
Joined 2005
Dan_Steele said:
Im using a 50k alps log for volume and 20K trimpot as sugested.
...
At the current levels (trimpot positions) as the volume is increased it sounds like one is tuning a radio in (that varying tone). When 12 oclock is reached all his disspears and the whloe system sounds perfect, when around 3 oclock is reached the hiss comes back for the last 1/4 turn. (however all hum in the mid/low amp diminishes)

For clarification, the trimpots are the ones in the buffer stage as shown in Figure 3 here http://www.sound.westhost.com/project09.htm, correct? And your volume pot connected between the output of the active crossover and the input of the amp, right?

If so, your 50k volume pot is not a good match for the 20k input impedance of the noninverting LM3875 circuit. It would be better if you used a lower value pot, 10k or lower if possible. You may also try grounding the case of the alps pot (if you haven't already) and wiring in a small cap across the inputs of the LM3875 as pinkmouse suggested.
 
Hi BWRX,

The trimpots are located on the board and connected before the output buffer as shown.

The 50K pot is not after the Xover, but before, therefore only a dual pot was required (would need a quad if on the output side)

The pot comes after a elma selector switch which has every other input shorted as suggested by morgan jones.

This lends itself to miniumum wiring lenghts in the preamp without the use of shaft externders.

I temporarily have a earth on the alps case, this prevents hum when the volume knob is touched

Im going to have a go at pinkmouses suggestions, just ordering the 220pf's now.

Regards

Dan
 
Ex-Moderator
Joined 2005
Hi Dan.

Yes, that makes more sense that the volume pot is at the input. I wasn't thinking that you'd need a 4 channel pot if it were on the output of crossover...

Well, don't worry about it's value then :) But do try grounding the metal part of the pot and adding the cap across the inputs of the LM3875.
 
I've ordered some polyprop's and also polycarbonate types.

Theres some 220pf silver micas on ebay at the moment but they looked a little 'vintage'

I steered clear of ceramic as I've can remember reading it has no place in an audio signal path (which I guess this application is along those lines)

Dan
 
Right Guys, an Update;

The 220pf polyprops arrived y/day I installed one to one of the amps and it changed all of them for the better (disconecting the one with the cap brought the hiss back)

So I then went on to install them in all four amps, I am pleased to say that the radio tuning / bizzare crackling has gone (all but on the last 10th of a turn on the pot, but im satisfied at the moment as it will never go that high)

There was still a low frequency buzz that was annoying me, after searching high and low, I decided to see what the amps are going to look like when in there finished position (ontop of the power supplies) and low and behold the vast majority of the buzz dissapeared!

I also tried the resistor and cap in the feedback loop to eliminate DC, but this made no audible difference.

Now there is still a hiss in the background, but only around 10% of what was there before, but I'm confident im moving in the right direction to compeltly get rid of it.

Here are a few pics I took last night (they were taken on my mobile mind)
Heres a pic of two of the amps

amp.jpg


And one of the X-over and PSU

activexover.jpg


This is the whole setup, showing the amps on top of the psu's
WholeSetup.jpg

Every PSU has a safety earth connected, however there is no connection between the amps and their chassis at present.
The cables between the amps and the PSU's are screened and connected to the safety earth.
The PSU's still need the rear feet attaching and so does the x-over. Also, I will make the earth connection shown with the croc clip permanent and change over two of the outputs. And trim down the pot / selector shafts to get a better fit with the knobs.
Finally, I've bought some 8mm screened cable for interconnects and will be making my own of the shortest length possible to hopefully further reduce the noise.


Finally

speaker.jpg


These will be finsihed soon and hopefully I will be able to sleep / see my girlfriend then!

All in all, not bad for a 19 year old if I dont say so myself :)


Thanks guys you've really helped me here! :worship:

Edit: Also had some screened IEC cables arrive yday so they may have added to the overall effect
Edit2: There is a chip shaped hole in the base of each amp so that the back of the chip is directly coupled to the heatsink with a dab of compound in between
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.