Hi- & low-pass filtering: how about Bhrngr PEQ2200/T1951

[clintrubber]

Looking for a simple freq-adjustable hi- & low-passfilter
I saw the Beh. PEQ2200 (for EU39...).
Apart from the 5 parametric bands it has a second order
hi- & low-pass.
It's adjustable from 10-400 Hz respect. 2k5 - 30kHz.

Intended application is for some soundshaping (lo-fi-FX)
and perhaps some sidechain-filtering.

I was originally intending to wire up a few opamps & dual-pots
myself, but for such prices it'll be clear the soldering iron is better
aimed at other things.

In addition, their the dual channel T-1951 (tubed,
but @ low voltage I thought, EU115...) has such a
filtering-possibility as well, but in a shelving manner.
Based on what I could find in the documentation I figure
this gives less steep filters.


So my question is twofold:

Anyone with experience with using freq-adjustable
low- & high-pass filters any idea if shelving-type of
filters are less suited for this task ?

Anyone any experience with the PEQ2200 or T1951 ?
I guess for these prices you can't complain - but at the
same time, if it's junk then better don't spend any
money on it at all.


Thanks,

Peter

 

[gyraf]

My favorite adjustable hi/low cut filter is the one in the Calrec EQ:

Very simple also.. "AOT" means select component (for good tracking)

Jakob E.

 

[clintrubber]

Thanks Jakob for the schematic, not too much soldering needed indeed.
Could wire one up & using normal log pots.
[edit: that's not true, the frequencies @ CW & CCW will be exchanged]

Should have some 3307x around - what were the original opamps ?

BTW, I also take that posted schematic as a suggestion to skip the Beh.-stuff :wink:

Bye,

Peter

 

[gyraf]

Originals were LF353-type opamps..

 

[gar381]

Hi Jakob

Thanks for posting this calrec schemo!! :grin:

Does A-log mean antilog (rev. log)??
Also..what would be a good AOT starting value
for C12??

Thanks :guinness:

GARY

 

[gyraf]

Does A-log mean antilog (rev. log)??

Nope, I believe it means standard log. The rev.log. ones are marked - rev.log..

Jakob E.

(I'm moving this thread into the lab..)

EDIT: 8 years later, jan.2014, we found the pots to be negative-logaritmic ("C"-curve) - see page2 of this thread

 

[jensenmann]

Jakob, which OPs are used in this filter?
:sam:
Jens

 

[gyraf]

Look up..!

 

[jensenmann]

....



:sam: for Jakob being very patient with some of us :wink:

Jens

 

[clintrubber]

Originals were LF353-type opamps..

Thanks Jakob.

FWIW (since more than a few opamps will do), that Calrec circuit could put a few LF351N opamps around here to good use (as I understand it these are the single channel versions of the '353).

 

[bcarso]

What are the notations with C12 and C19? Are those trim caps?

 

[clintrubber]

See above: Adjust On Test.
(

Very simple also.. "AOT" means select component (for good tracking)

)

Dunno about their starting value though, but that'll soon be clear when some brainpower or a calculator or simulator is thrown at the circuit. :wink:

 

[bcarso]

Ooops there it is---thanks I should have read that.

Are these supposed to be butterworth, ie.e, Q of .707? That would be the traditional assumption. I would have done the equal R equal C version with a gain of 1.585, but to each his/her own.

 

[clintrubber]

[quote author="bcarso"]Ooops there it is---thanks I should have read that.

Are these supposed to be butterworth, ie.e, Q of .707? That would be the traditional assumption. I would have done the equal R equal C version with a gain of 1.585, but to each his/her own.[/quote]

A thread from a while back, but...

Thank for the response Brad.
Time to dive again into some filter-math, but I guess you're suggesting a Bessel (Thomson) with that gain of 1.585, correct ?


Regards,

Peter

 

[tony dB]

I had 6 of those beh EQ's (old model) before I went digital in '96.
They sounded not bad and got me the (heavy processed for modern metalmusic) kicks and snaresounds very easy. I asked a few of my clients (EE's) if they could improve on these as a lot of the components were carbon resitors and not topquality opamps etc... Pots were also a little to much 'el cheapo', so there was pimpability. When measured for noise, phase, etc... the measuring results were not good at all, although to me it did not sound that bad as what we measured.
Most of the reactions I got were don't mess with it... but propably, that was because non of these guys really wanted to dig into it I think. Sold them all for redicously little cash when upgrading my studios back then.

 

[recnsci]

Im not Brad but that is butterworth he is sugesting.
Filter math is super easy with equal valued Rs and Cs.
cutoff is at f=1/(2piRC) and Q=(3-A)/2 where A is gain of that
noninverting stage. Note that passband gain is A.
My suggestions for A would be 1, 1.5 and 2 . If you have dual three
position switch you could have compensation of passband gain
with attenuator on output. Or use another opamp as unity buffer and
feed it from - input of first opamp.
Also, low Ibias or fetinput opamp would be nice idea.

cheerz
urosh

 

[clintrubber]

[quote author="recnsci"]Im not Brad but that is butterworth he is sugesting.
Filter math is super easy with equal valued Rs and Cs.
cutoff is at f=1/(2piRC) and Q=(3-A)/2 where A is gain of that
noninverting stage. Note that passband gain is A.
My suggestions for A would be 1, 1.5 and 2 . If you have dual three
position switch you could have compensation of passband gain
with attenuator on output. Or use another opamp as unity buffer and
feed it from - input of first opamp.
Also, low Ibias or fetinput opamp would be nice idea.

cheerz
urosh[/quote]

Thanks Urosh,

Hmmm, I saw a Q-formula for Sallen-Key with equal R & C stated as Q = 1/(3-A), but I should first have a closer look to some files before we take that statement from me too serious.

Your suggestions are nice, I recall that that '-'-output method, nice one.

The thing is, I'm actually curious about the why of the various circuits from different manuafacturers. I've seen variable HPFs & LPFs from Calreac, Porter, D&R, ...) Some do the equal R & C, others do different values.
I recall also a Krohn Hite filter that has switchable Bessel/Butterworth.


So it's not difficult to go really overboard with switches etc, while a simple circuit with only switches for bypass would be just enough ! :wink:

Regards,

Peter

 

[recnsci]

well you are right about Q, I was writing straight out of my head
(with out actually using that part of my body to think for few seconds).
Calrec did miss matched Cs in HPF so they could use standard 10k
Rs around opamp for butterworth response. Or least I think it was reason.
If you dont want switchable Q my absolute favourites are exactly
what calrec used: 0.7 for HPF and 1 for LPF. I use that 9 times out of 10
when I filter.
And SK is just about best topology for this stuf in my world: no DC across pots if opamp has
low Ibias, and use of single opamp that doesnt need to be stellar and isnt
loaded too heavy. And it lends it self easyly to 2-3 transistor discrete
implementation.

cheerz

 

[clintrubber]

[quote author="tony dB"]I had 6 of those beh EQ's (old model) before I went digital in '96.
Sold them all for redicously little cash when upgrading my studios back then.[/quote]
Hmm, if those older models were already a bit flimsy then
I bet the people that bought them from you for little cash got a better deal than those people that couldn't withstand the EU39 for the newer ones from Thomann :thumb:

 

[clintrubber]

[quote author="recnsci"]well you are right about Q, I was writing straight out of my head
(with out actually using that part of my body to think for few seconds).
Calrec did miss matched Cs in HPF so they could use standard 10k
Rs around opamp for butterworth response. Or least I think it was reason.
If you dont want switchable Q my absolute favourites are exactly
what calrec used: 0.7 for HPF and 1 for LPF. I use that 9 times out of 10
when I filter.[/quote]
Hmm, why different for HPF & LPF ?

And SK is just about best topology for this stuf in my world:
cheerz

I've had a very short look at Multiple FeedBack-topologies, as reported lower THD-possible as compared to SK, but it won't be as easy to make it variable.So let's skip, if some topologies are good enough for Calrec etc then it'll be like that for me as well.

no DC across pots if opamp has
low Ibias, and use of single opamp that doesnt need to be stellar and isnt
loaded too heavy. And it lends it self easyly to 2-3 transistor discrete
implementation.

I've seem mentioned or used: LF353, TL07X, NE5532. I'd like to use the '5532/5534, but am not sure it'll be good enough w.r.t. avoiding scratchy pots.

and use of single opamp that doesnt need to be stellar and isnt
loaded too heavy

I'll post the D&R-circuit tonight, they use TL074 but 'help' the opamp a bit. And they avoid loading below 2k7, which is fine enough IIRIC.

The discrete implementation (say as a CFP with some gain) has crossed my mind as well, but that'll cost some more time to DIY; I better finish this more quickly.


I still do want to add a SPDT per filter though, smartly placed this can switch between 1st & 2nd order.

The Q-switch sounds good as well (the active electronics in Alembic basses has such a switch for instance, 3 position).

Not sure about a Butterworth/Bessel-switch. Anyone actually seen desk-EQs with Bessel response ? Too mellow ?

It'll have a bypass-switch of course; I don't want to have to set the freq-knobs at their extremes to get the circuit non-filtering and the switch makes comparisons easier. (I realize it might require a make-up control for fair comparisons.)

Bye !

Peter