Line level crossover

Design and use of the various types.
reVintage
Posts: 365
Joined: Fri Sep 07, 2007 9:54 am
Location: Sweden

Post by reVintage »

HI Pierre,
Tried to simplify a little. Have only made some guesswork about crossover-frequencies.
Attachments
prazza3way.PNG
prazza3way.PNG (64.82 KiB) Viewed 31711 times
Linelevel Xover 04.asc
(5.9 KiB) Downloaded 421 times
Brgds
Lars
prazza1
Posts: 38
Joined: Tue Nov 29, 2005 12:17 pm
Location: France

Post by prazza1 »

Thank you Lars,

The project starts to be feasible. I like your work. About frequencies it was only as an instance. They will have to be adjusted later.

Dave, what do you think about it?

Rgds, Pierre.
dave slagle
Posts: 2085
Joined: Sat Jan 22, 2005 3:54 am
Location: NYC
Contact:

Post by dave slagle »

I think the idea is a good one. It does have a lot of iron in it so I guess I have to like it :-)

How do you plan on level matching?

dave
reVintage
Posts: 365
Joined: Fri Sep 07, 2007 9:54 am
Location: Sweden

Post by reVintage »

Why not do like Gianluca and and substitute the filterloading resistors with pots?
Brgds
Lars
prazza1
Posts: 38
Joined: Tue Nov 29, 2005 12:17 pm
Location: France

Post by prazza1 »

Why not do like Gianluca and and substitute the filterloading resistors with pots?
Cause amps input are very low so I need very very low output buffers. The output buffer follower can provide this.
Maybe we can use pots instead of filter resistor (the 4.7 k) ?

Rgds, Pierre.
dave slagle
Posts: 2085
Joined: Sat Jan 22, 2005 3:54 am
Location: NYC
Contact:

Post by dave slagle »

Pots in place of the 4K7 should do just fine.

dave
reVintage
Posts: 365
Joined: Fri Sep 07, 2007 9:54 am
Location: Sweden

Post by reVintage »

It might be some noise when due to the DC base-current through the pots but they shouldn´t be moved so often. What about resistance? Lower values gives smaller chokes.....
Brgds
Lars
dave slagle
Posts: 2085
Joined: Sat Jan 22, 2005 3:54 am
Location: NYC
Contact:

Post by dave slagle »

I would expect a simple 12 step ladder attenuator would be just fine. Going to a lower Value indeed means smaller chokes but it also means bigger caps.

The output iron could also be tweaked so it is better tailored for the frequencies in question.

dave
prazza1
Posts: 38
Joined: Tue Nov 29, 2005 12:17 pm
Location: France

Post by prazza1 »

Dave,

I'd like to test step by step.
Maybe, first, I can test (and hear) it as a preamp, without crossover. So how about output trany ? How can we do ?

Thank you.

Regards,
Pierre
dave slagle
Posts: 2085
Joined: Sat Jan 22, 2005 3:54 am
Location: NYC
Contact:

Post by dave slagle »

do you mean a 1:1 bifilar with 50hy's of inductance as drawn in the sims above?

dave
prazza1
Posts: 38
Joined: Tue Nov 29, 2005 12:17 pm
Location: France

Post by prazza1 »

do you mean a 1:1 bifilar with 50hy's of inductance as drawn in the sims above?
This transformer was drawn for simulation purpose and it seems to work. Here I speak to the specialist: maybe can you make better ? If you think the above transformer is the best compromise (because all is compromise) so yes, the same. If no, I let you do and suggest. Than you.

Regards.

Pierre.
dave slagle
Posts: 2085
Joined: Sat Jan 22, 2005 3:54 am
Location: NYC
Contact:

Post by dave slagle »

I ran the DC operating point and the current is 30ma. While I can do a 50hy 30ma 300R device, it will be on a larger core than ultimately needed and since these will ultimately be for specific frequencies, I see no need for them to all be the same.

I can do the 50hy @ 30ma if the DCR's go up to 600 ohms per winding or get closer to the 300 ohm number and give 30hy's @ 30ma. I like the idea of going for the lower inductance but let see if Lars has an opinion.

Ultimately you could go with something like 50hy 600R for the bass 30hy 300 ohms for the mids and 15hy 150R (or use a smaller 80% nickel core) for the highs). My sims still show 30hy's in this situation being -1dB at 3hz so i'm not concerned :-)

dave
reVintage
Posts: 365
Joined: Fri Sep 07, 2007 9:54 am
Location: Sweden

Post by reVintage »

I choose 300ohm DCR as this would result in ca 600ohm Zout. With the signallevels we are talking about here together with Iq, smaller inductaces shouldn be a problem. Can do some sims of the harmonics if you want to.
Brgds
Lars
prazza1
Posts: 38
Joined: Tue Nov 29, 2005 12:17 pm
Location: France

Post by prazza1 »

Can do some sims of the harmonics
Yes, if you want. Thank you.
I choose 300ohm DCR as this would result in ca 600ohm Zout
Why 600 ? Don't we see through the trany 1:1 the transitor follower output impedance ?

Regards

Pierre
reVintage
Posts: 365
Joined: Fri Sep 07, 2007 9:54 am
Location: Sweden

Post by reVintage »

You must add the DCRs to the Zout of the emitterfollower. So it will still be ca 600ohm.
Brgds
Lars
Romy The Cat
Posts: 22
Joined: Fri Feb 13, 2009 1:52 am

Post by Romy The Cat »

reVintage wrote:HI Pierre,
Tried to simplify a little. Have only made some guesswork about crossover-frequencies.
reVintage.

The guesswork about crossover-frequencies at line level be a pin in ass, you would need some like of prototype type of crossover that you might lend in your local pro audio shop… I did not read the thread but I do not like what I see in the picture.

HF: get rid of that C1 and do for RL filter. It is not even arguable in my book

MF: This band passes works but if you have the top knee relatively high or if you objectives are too refined then it will not and you need to convert it into a series resistor and cap and coil to ground. It will give you the same filter but you will have no reactive complement in signal bypass

LF: If you cat off is low than you will have a large inductance with a lot of wire in it. Since it bass you would need a very low DCR that means think wire and very large size. Well, did you try to hide a large size coil inside of an amp? Good luck with that! So, you might want to use the same concept: no reactive elements in series application, or to use a resistor and a cap to ground. You might not like caps but THIS cap will not be a problem as it will only pass the shuntable signal that is NOT in the pass of the channel.

Rgs, Romy the Cat
reVintage
Posts: 365
Joined: Fri Sep 07, 2007 9:54 am
Location: Sweden

Post by reVintage »

Romy,

Come back when you have read the whole thread! And don´t give me the advice, better you address your ideas to Prazza or Dave.

As I understand it the thread is primary about Prazza wanting to mate transformers with transistors in a filter of his own design.
Brgds
Lars
Romy The Cat
Posts: 22
Joined: Fri Feb 13, 2009 1:52 am

Post by Romy The Cat »

reVintage wrote:Romy,

Come back when you have read the whole thread! And don´t give me the advice, better you address your ideas to Prazza or Dave.

As I understand it the thread is primary about Prazza wanting to mate transformers with transistors in a filter of his own design.
ReVintage,

I did not give an adviser to YOU but I posted it a public forum. For you it cost as much as you invest into reading my post and if you did not “get” my objectives then it is fine with me. I still did not read the thread and I posted what I posted based upon your image - if the image has nothing to do with content of the thread then what does it do there? The point of all of it is that I spent a LOT of time experimenting, analyzing and listening different line-level crossover topologies in different applications. I have to throw into the mix the proposal that in contrary to many others “soldering” audio people I know sound and I know what to listen while I am listening. If you feel annoyed that I have expressed my observation about the filtering schema proposed in your drawing then discard what I expressed and forget about it – I truly have no need to convince you in anything.

The Cat
prazza1
Posts: 38
Joined: Tue Nov 29, 2005 12:17 pm
Location: France

Post by prazza1 »

Romy and ReVintage, please keep cool. I'm ready to accept any advice.
In the meantime I had worked a little on filter and I have made something a little like Romy suggestion.

Romy, any critic ?

ReVintage, thank you for your advices too, I appreciate a lot to work with you and Dave because I have not your experience. If you could give us results about harmonics simulation.. After I will see with Dave about continuation.

Thank you for all, you people !

Regards,
Pierre
Attachments
Filter.jpg
Filter.jpg (66.42 KiB) Viewed 31349 times
reVintage
Posts: 365
Joined: Fri Sep 07, 2007 9:54 am
Location: Sweden

Post by reVintage »

Hey Prazza,
The network loads the driver harder at low frequencies. It will probably be better if you go for as in this case 4,7k all the way to get equal loading independant of frequency.

Personally I would go for lower series-resitances all over. Your driver could easily take something in the ballpark of 2k.
Brgds
Lars
Romy The Cat
Posts: 22
Joined: Fri Feb 13, 2009 1:52 am

Post by Romy The Cat »

Pierre, this is a bit tricky.

The filters look now all right to me, presumably that the values are right. You need to look for the environmental damage. I do not know SS and I have no idea if your transistors will do OK with a cap or inductor sitting right at their base. I do not know if your driver will be able to drive this low impedance. So, you need to look into all of the environmental variables.

What is tricky here is to find out if the different load for different channels will be OK. Of course if you make then the same then you out of muddy water BUT I have seen situation what moderation of LF channel impedance in relation to other channels did make an interning effect. Your MF channels will be no changed but your LF channel will or might. Basically you moderate current that will be reassigned between the channels. In theory current does nothing in this case and you care just about voltage. In practice, if you have very capable LF section, then you can recognize the difference. Back in 2001 I have a 2 pairs of ML2 and I was driving my Bass ML2 with twice higher or lower impedance (I do remember already). It was very auditable difference. Well, in my case I drover cable, so high current “pressed” them harder, that is always better. In your case it might be not a factor – you need to try it. Also, the large amount of magnetics might have some masking effects to your bass quality, so it might be all irrelevant in your case.

The Cat
reVintage
Posts: 365
Joined: Fri Sep 07, 2007 9:54 am
Location: Sweden

Post by reVintage »

Hi Prazza,

Check your filter once again :wink: ! If you buy the concept with a 6H choke picking up hum when in the bass-channel it will pick-up equally much hum when situated in the midchannel.

One shouldn´t mix up what is going on inside the buffered filter with what is happening from the output of the filter to the input of the poweramps. For the input-buffer it is important to see the same load over the entire frequency-spectrum!
Brgds
Lars
prazza1
Posts: 38
Joined: Tue Nov 29, 2005 12:17 pm
Location: France

Post by prazza1 »

Dave,

I would like to try first 30H @ 30mA transformer in line level mode before to pursue with filter, to hear the result.
We had taken 300 ohms DCR for simulation. Can't we have very less ? Where can we reasonably go?

Brgs,

Pierre
dave slagle
Posts: 2085
Joined: Sat Jan 22, 2005 3:54 am
Location: NYC
Contact:

Post by dave slagle »

Hello Pierre,

The 30hy 30ma was a guess based on using the same wire gauge on the primary and the secondary. This may have been covered but do you want a lower DCR for biasing reasons?

You can think of it this way, both the primary and secondary appear in series for the output impedance so you will have a ~600 ohm resistive Zout (300 + 300 assuming the sand is near 0). nothing says the wire cant be chosen to get a 150 ohm primary but then you end up with a 450 ohm secondary. Essentially the DCR is set by the winding window size and the number of turns. There are two ways to lower the overall DCR and that is a bigger window (larger core which increases $$) or fewer turns which requires the choice of lower inductance / current.

At this point there are just ballpark guesses (+- say 20%) and I'd have to wind and measure a sample to get accurate numbers.

dave
prazza1
Posts: 38
Joined: Tue Nov 29, 2005 12:17 pm
Location: France

Post by prazza1 »

Dave,

I asked that to have lower output impedance not for bias (this is ok). I already have a preamp with OPT and I have measured DCR of secondary around 8 ohms. But maybe this OPT is used in Parallel Feed Single-Ended (with capacitor) mode so maybe with low primary and secondary DCR. Your opinion ?

Pierre
dave slagle
Posts: 2085
Joined: Sat Jan 22, 2005 3:54 am
Location: NYC
Contact:

Post by dave slagle »

I'm sure the 8 ohms is the result of a step down. The easiest way to get the DCR down is to reduce the inductance a bit. Do you really need 30hy's?

dave
prazza1
Posts: 38
Joined: Tue Nov 29, 2005 12:17 pm
Location: France

Post by prazza1 »

I think no. What can be minimal load with lower inductance ?

Pierre.
dave slagle
Posts: 2085
Joined: Sat Jan 22, 2005 3:54 am
Location: NYC
Contact:

Post by dave slagle »

I ran the step command in spice to look at the inductances and dropping it to 3hy only makes a slight visible change so how about something like 6hy to be "safe" If we come up with a realistic inductance that you are happy with then i'll see what the DCR will be.

dave
Attachments
Linelevel Xover 04 step.asc
step commands to change inductance added.
(6.18 KiB) Downloaded 363 times
The blue line is barely visible just below the green one.
The blue line is barely visible just below the green one.
Picture 3.png (26.4 KiB) Viewed 31063 times
prazza1
Posts: 38
Joined: Tue Nov 29, 2005 12:17 pm
Location: France

Post by prazza1 »

how about something like 6hy to be "safe"
Yes, we can try.

Pierre
dave slagle
Posts: 2085
Joined: Sat Jan 22, 2005 3:54 am
Location: NYC
Contact:

Post by dave slagle »

Hi Pierre,

Sorry for being slow in my response. I am at the end of my stock of the core material needed and the scheduled delivery is the end of July. I hope you can wait until then to continue. (please feel free to remind me)

dave
Post Reply