Hi all.
I recently got a pair of S800 passive. They are in great condition and was a steal so i could not resist.
This is not exactly the one i would have took as i prefered their bigger sisters (S1200) when i worked with System but really they were a bargain.
Overall i'm satisfied with them given their limitations ( 8" is the lower limit to me, smaller diameter is too high fc for my taste and the smaller models don't have the output capability i sometimes need spl wise).
But there is something which bother me around 1,5khz. I couldn't explain clearly what it is but they sound different that their active twins which doesn't have that thing that bother me ( i've worked with both and know them well).
So i searched for the s800a schematic and studied it as i plan to modify my passive to something else ( running them as the mid/high of a 3 way, closing the ports to make them sealed and relieve the 8" from bass duties with dedicated sealed sub at around 250/300hz).
And as Tannoy did a good job ( at least i like them ) on the filter of the active version so why reinvent the wheel?
There is difference between them: on the active, fc is lowered 200hz to 1,6khz and there is some allpass used to compensate for the different emmissing point on the woofer and tweeter.
Initially i was surprised to see the allpass filters used on the woofers but second thoughts it make sense, the tweeter is located behind the woofer so the later is in 'advance' refered to the tweeter.
Ok but what puzzle me is that it use two all pass in series one 'leading phase' ( ic3d) then one 'lagging phase' (ic2b) and both have the same fc of 4420hz.
So here is my question what is this part of the circuit doing exactly? Aren't both leading and lagging allpass in series canceling each others? And how could i estimate the delay time induced ( as i suppose there is one)?
Thank you for answers.
(Schematic attached).
I recently got a pair of S800 passive. They are in great condition and was a steal so i could not resist.
This is not exactly the one i would have took as i prefered their bigger sisters (S1200) when i worked with System but really they were a bargain.
Overall i'm satisfied with them given their limitations ( 8" is the lower limit to me, smaller diameter is too high fc for my taste and the smaller models don't have the output capability i sometimes need spl wise).
But there is something which bother me around 1,5khz. I couldn't explain clearly what it is but they sound different that their active twins which doesn't have that thing that bother me ( i've worked with both and know them well).
So i searched for the s800a schematic and studied it as i plan to modify my passive to something else ( running them as the mid/high of a 3 way, closing the ports to make them sealed and relieve the 8" from bass duties with dedicated sealed sub at around 250/300hz).
And as Tannoy did a good job ( at least i like them ) on the filter of the active version so why reinvent the wheel?
There is difference between them: on the active, fc is lowered 200hz to 1,6khz and there is some allpass used to compensate for the different emmissing point on the woofer and tweeter.
Initially i was surprised to see the allpass filters used on the woofers but second thoughts it make sense, the tweeter is located behind the woofer so the later is in 'advance' refered to the tweeter.
Ok but what puzzle me is that it use two all pass in series one 'leading phase' ( ic3d) then one 'lagging phase' (ic2b) and both have the same fc of 4420hz.
So here is my question what is this part of the circuit doing exactly? Aren't both leading and lagging allpass in series canceling each others? And how could i estimate the delay time induced ( as i suppose there is one)?
Thank you for answers.
(Schematic attached).
An allpass filter changes the phase by 90°. So two in series change the signal by 180°, which means, it's the same phase as the original signal. The difference is, a sine wave looks the same with reverse polarity but an impulse is delayed double the time of a single allpass and each adds another order to the filter characteristics.
Hi Icg,
First thank you for your answer!
Well yes i agree about your definition and it was my initial thoughts too but when i studyed the schematics i was astonished to see both variations of allpass used in series and at the same frequency: the 'leading' variation which is supposed to make signal appearing 'before' the initial signal and with inverse phase refered to the 'lagging' one which make the signal appearing 'after' the initial signal.
If both allpass cells have been the same 'lagging' ( or 'leading') variations your answer would have been ok, but in this case this isn't the case ( well i may not understand well and mix things up but i already spent 2 weeks trying to figure the action of the circuit and this is definitely not a case of doubling the same 'cell' to double the order of the allpass 'cell').
The other thing that puzzle me is the fact that they use same fc for both 'cells': the delay is supposed to be effective at lowpass fc of 1600hz ( ic3c) and from my understanding of the allpass behavior they may cancels each others. But as the effect should be relevant at 1/3 the allpass fc maybe something happens i don't grab? And i don't see Tannoy spending money on parts which counteracts themself ( it is not rational from an economical point). Maybe this have to do with the notch implemented by ic3a/b?
This make me scratch my head!
The only reference i've seen of both variations are there and as pointed by R. ELLIOTT, the behavior of both cells should be different:
Active Filters
And this is where the article is frustrating as it stop where i need it to dig deeper!
That said i could just implement it and test it latter but i like to know what i'm doing and why.
First thank you for your answer!
Well yes i agree about your definition and it was my initial thoughts too but when i studyed the schematics i was astonished to see both variations of allpass used in series and at the same frequency: the 'leading' variation which is supposed to make signal appearing 'before' the initial signal and with inverse phase refered to the 'lagging' one which make the signal appearing 'after' the initial signal.
If both allpass cells have been the same 'lagging' ( or 'leading') variations your answer would have been ok, but in this case this isn't the case ( well i may not understand well and mix things up but i already spent 2 weeks trying to figure the action of the circuit and this is definitely not a case of doubling the same 'cell' to double the order of the allpass 'cell').
The other thing that puzzle me is the fact that they use same fc for both 'cells': the delay is supposed to be effective at lowpass fc of 1600hz ( ic3c) and from my understanding of the allpass behavior they may cancels each others. But as the effect should be relevant at 1/3 the allpass fc maybe something happens i don't grab? And i don't see Tannoy spending money on parts which counteracts themself ( it is not rational from an economical point). Maybe this have to do with the notch implemented by ic3a/b?
This make me scratch my head!
The only reference i've seen of both variations are there and as pointed by R. ELLIOTT, the behavior of both cells should be different:
Active Filters
And this is where the article is frustrating as it stop where i need it to dig deeper!
That said i could just implement it and test it latter but i like to know what i'm doing and why.
Last edited:
To make the signal appear at the output before the input is not possible. Yes, I've read the article. They specifically determine "for a signal that lasts more than a few cycles it really does happen." Music isn't sine waves, so that doesn't apply because it's not swung in. You can see at graph 7.6 in your article it needs time (one or even more cycle(s)) to swing in.
Even if it did apply, it would be indistinguishable by the ear as the sound would be exactly the same as a cycle further, it doesn't 'know' to which cycle it actually belongs as a steady tone got no time information in it opposed to an impulse or a series of impulses where the delay to each other is audible.
Well, why don't you measure what it exactly does? You don't even need a 2 ch digital osci, you can do that with a 2 channel measurement with your soundcard, the voltage is line level. There are also simulators for it (haven't used one in years, can't recommend any, sry)
Allpass Filters
More General Allpass Filters
(No, I didn't pull out the calculator or a simulator to check it)
Well, I'm not so sure now, I'm thinking about a bridged t delay filter. I can't concentrate enough right now and later my lazyness might kick in..
Well, if you want to know it, you know why and you know what you have to do because of the schematics.
Got a passive pair as well. Hit similar issue. A sweep through the crossover region revealed that the tweeter and bass were modulating due to poor phase matching. Tannoy screwed up. Need to redesign xover.
Well, in this case, since he already got the active speakers (with the amps) I'd suggest to use a dsp instead of re-modeling the setup with discrete filters and change the board over and over.
Icg,
Thank you for thoughts.
To be clear i don't own an active but used to work with in one studio where i did some works so i know them and how the mix i've done using them sounded and when playing back on the passive their is something different happening ( and this isn't amplifier related ( Amcron/Crown D150) and i use calibrated level when working so i can reproduce conditions almost exactly - acoustic is the parameter i don't have hands on but i would say that my domestic conditions are 'better' that the one of that specific control room, my room is bigger -almost 200m3- with sloped ceiling).
Whatever i will use the filter Tannoy developed for the Active version, it is not really complicated to implement, but as i stated i prefer to understand things before duplicating them.
In the end i may just populate a protoboard and measure what result it give...
That said i have a (very) nice dsp processor for my threeways and could use it to 'clone' Tannoy's solution but i need to know delay time used rather to guess, go by trial and error or measure things... as i'm lazy and lacks spare time.
Spladski,
Yes i've heard what i think is 'modulation' on some tracks too. I don't know if it come from the filter though, i think this is more related to music genre ( very low dynamic source material with lot of subs) and limitation of 8".
Hence my wish to limit low end duties on the woofer and go sealed to limit mid leakage( and use of linkwitz transform for improved transient response). But i may be wrong about that too...
I would not say Tannoy screwed things with the system 800 passive though. I'm regularly blown away by the 'transparency' they have, great recordings sound great, poor recording sound poor... my own work is displayed honestly and i recognise room and instruments characteristics easily even at low level spl. Even my own mystake or bad choices are there!
I just think there is room for improvement which could be rewarding.
Thank you for thoughts.
To be clear i don't own an active but used to work with in one studio where i did some works so i know them and how the mix i've done using them sounded and when playing back on the passive their is something different happening ( and this isn't amplifier related ( Amcron/Crown D150) and i use calibrated level when working so i can reproduce conditions almost exactly - acoustic is the parameter i don't have hands on but i would say that my domestic conditions are 'better' that the one of that specific control room, my room is bigger -almost 200m3- with sloped ceiling).
Whatever i will use the filter Tannoy developed for the Active version, it is not really complicated to implement, but as i stated i prefer to understand things before duplicating them.
In the end i may just populate a protoboard and measure what result it give...
That said i have a (very) nice dsp processor for my threeways and could use it to 'clone' Tannoy's solution but i need to know delay time used rather to guess, go by trial and error or measure things... as i'm lazy and lacks spare time.
Spladski,
Yes i've heard what i think is 'modulation' on some tracks too. I don't know if it come from the filter though, i think this is more related to music genre ( very low dynamic source material with lot of subs) and limitation of 8".
Hence my wish to limit low end duties on the woofer and go sealed to limit mid leakage( and use of linkwitz transform for improved transient response). But i may be wrong about that too...
I would not say Tannoy screwed things with the system 800 passive though. I'm regularly blown away by the 'transparency' they have, great recordings sound great, poor recording sound poor... my own work is displayed honestly and i recognise room and instruments characteristics easily even at low level spl. Even my own mystake or bad choices are there!
I just think there is room for improvement which could be rewarding.
Last edited:
Oh, I'm sorry I presumed the wrong circumstances. Well, if you'd use class D amplifier boards or LM3668 based amps to go active for a quite reasonable price. But you probably already planned with some amps you already got or have in mind buying/building.
The different room will change the reproducton conditions very much. The room size pretty much asks for bigger speakers unless you're only using them in the near field (what they are designed for). It will still sound different. A sloped ceiling can be advantagous or pose a serious deficite, depending on angle, height etc. The room impression will change too outside the near field, the sound of a speaker changes a lot with a different enviroment and placement.
Whatever i will use the filter Tannoy developed for the Active version, it is not really complicated to implement, but as i stated i prefer to understand things before duplicating them.
In the end i may just populate a protoboard and measure what result it give...
That said i have a (very) nice dsp processor for my threeways and could use it to 'clone' Tannoy's solution but i need to know delay time used rather to guess, go by trial and error or measure things... as i'm lazy and lacks spare time.
You don't have to replicate the active crossover to get the delay, you can feed the speaker electronics a needle impulse and measure the delay at the amp outputs or wires of the driver with a 2 ch measurement (stereo of a soundcard works fine).
With a DSP you can improve the speakers in a great scale but they won't sound exactly like the other pair and the setup requires measurement (there are good, cheap measurement mics and free software, ie. ARTA) and effort and time to build the setup.
Icg,
Yes i am a man with a plan!
Class D amp or 7293 like the original unit. Already found some board which fullfil requirement (90w/8 Ohms from connexelectronics or tda7293 stand alone units). If using the class D board ( which is a litlle bit too much powerful vs the original) the only thing to keep in mind is to add a -4db pad in front of the tweeter dedicated board ( the overall gain of the original chipamp for tweeter is lowered by 4 db by component choice around the tda 7293).
For rest of electronic, i want to duplicate the schematic except the low end 'manipulation' or 'high shelf' +2db compensation i don't need ( well i could but prefer something with more fine tuning. I've already spoted a dedicated circuit from S. Dove which does same thing with more finesse).
About room acoustic, you don't have to convince me be sure!
Let's say that my room is almost three time bigger than the control room i used the S800a. The cathedral ceiling in my living room does have an angle of more than 30* ( more like 40*) starting from 2,3m on the front wall to 3.9m at the higher spot in the room, in width it is 5,7 meter wide, in depth 6.7m.
The room has a leacky roof and i have a door which open on the rest of house too if i want ( read almost no complicated room mode in the very low). Really i'm lucky, and my big threeway looks tiny in there and never sounded this good! ( 15"+3"+1", 750hz/4,5khz initial passive modified to 500hz/6khz FIR thanks to linkwitz transform on mid and use of FIR filters on dsp processor).
Compared to the 2,5m height of the control room and if you make a bit of ray tracing (including +/- 15* to mimic reality rather than simplistic approximation) you'll see why i think my place is 'better' ( in spite of being non treated at the moment... too long a reverb time). And i don't have a 3m x 1,2m wide reflector in front of me neither ( but sometimes i miss that! But not for the acoustic properties nor the 'room heater' characteristics of big analog desks! ).
Yes i could take measurements and rebuild a new crossover, but as i said i'm lazy... and liked the active version of s800. I will just duplicate it for the moment. If i find the urge to change them latter then yes, i will rebuild a box, use the drivers in it and go for a full new thing using the Lake FIR filter capability ( at 1,6khz the delay induced is lower than 2ms which is tolerable for most musicians in case of monitoring and tracking using them). But for now they will still be some small box i can easily move.
Yes i am a man with a plan!
Class D amp or 7293 like the original unit. Already found some board which fullfil requirement (90w/8 Ohms from connexelectronics or tda7293 stand alone units). If using the class D board ( which is a litlle bit too much powerful vs the original) the only thing to keep in mind is to add a -4db pad in front of the tweeter dedicated board ( the overall gain of the original chipamp for tweeter is lowered by 4 db by component choice around the tda 7293).
For rest of electronic, i want to duplicate the schematic except the low end 'manipulation' or 'high shelf' +2db compensation i don't need ( well i could but prefer something with more fine tuning. I've already spoted a dedicated circuit from S. Dove which does same thing with more finesse).
About room acoustic, you don't have to convince me be sure!
Let's say that my room is almost three time bigger than the control room i used the S800a. The cathedral ceiling in my living room does have an angle of more than 30* ( more like 40*) starting from 2,3m on the front wall to 3.9m at the higher spot in the room, in width it is 5,7 meter wide, in depth 6.7m.
The room has a leacky roof and i have a door which open on the rest of house too if i want ( read almost no complicated room mode in the very low). Really i'm lucky, and my big threeway looks tiny in there and never sounded this good! ( 15"+3"+1", 750hz/4,5khz initial passive modified to 500hz/6khz FIR thanks to linkwitz transform on mid and use of FIR filters on dsp processor).
Compared to the 2,5m height of the control room and if you make a bit of ray tracing (including +/- 15* to mimic reality rather than simplistic approximation) you'll see why i think my place is 'better' ( in spite of being non treated at the moment... too long a reverb time). And i don't have a 3m x 1,2m wide reflector in front of me neither ( but sometimes i miss that! But not for the acoustic properties nor the 'room heater' characteristics of big analog desks!
). Yes i could take measurements and rebuild a new crossover, but as i said i'm lazy... and liked the active version of s800. I will just duplicate it for the moment. If i find the urge to change them latter then yes, i will rebuild a box, use the drivers in it and go for a full new thing using the Lake FIR filter capability ( at 1,6khz the delay induced is lower than 2ms which is tolerable for most musicians in case of monitoring and tracking using them). But for now they will still be some small box i can easily move.
I have several Tannoy models. The Eaton, Saturn, DC1, System 800. I have also serviced their older larger monitors. Tannoy did go through a period where they lost their way in my view. The models are not consistent in performance. The Saturn and the System 800 have the worst xovers. I completely redesigned the Saturn xover which took a lot of effort. Although the Saturn and Sytem 800 have 8 inch drivers, they behave differently. The System 800 driver doesn't go as low. I fixed the modulation issue with the 800 by modding the xover. However, the response is no longer flat. I was a bit disgusted with it, and relegated it to a DJ monitor. The effort to fix it properly was too great. The others I can use for mixing.
Hi Spladski,
Yes it seems this 8" is not easy to cross using passive. That is why i've choosen to try go active with them and hence my question about the allpass cells.
I've made some progress about this. I've got a layout sorted for the circuit on protoboard. Need to triple check it though and i could improve some things i'm not totally satisfied with in component placement and ground layout too. I'll probably use a starground strategy.
I'll probably use 2134 in place of tl7x. And maybe a That 1206 in place of 275 as line input receiver, this will be determined when starting build.
About the allpass cells, well ICG's first answer to my question may well be what happen in the circuit: i've probably didn't understood R.Elliott's article and expected something more complicated than it is in reality. Anyway, i asked for help to understand this in another (great) forum and didn't have any answer.
I thought knowledge about allpass were common with experienced designers and electronicians but this doesn't seem to be the case.
I'll take some measurement to have answers once the circuit is running.
Spladski you may be right about the last series of Tannoy's pro monitors. I've seen some comments of users really disappointed about the 800. Maybe they had quality control issues, maybe this is by design but it seems their is indeed inconsistancy between S800 that is sure.
Mine are dark grey not navy blue, and the last digit is different in serial between the 2 (213103p /213102p).
I won't say i could not work on them neither that i don't enjoy listening to them, they are quite revealing in my view. Except from this thing in 1,5k ( and that i prefer closed box behavior overall) i'm quite pleased. Maybe i'm lucky and got a nice pair?
I've spent some time with 1200 too and fell in love with them. As already said liked the Active 800. Not convinced by the 600 either passive or active but it could be 'bias expectations' from my side.
I've heard and worked with Dmt 2 too. 12" and 15". Liked them both especially the 15" but they were inwall mounted and think were active filtered and biamped iirc. Not a typical use.
Yes it seems this 8" is not easy to cross using passive. That is why i've choosen to try go active with them and hence my question about the allpass cells.
I've made some progress about this. I've got a layout sorted for the circuit on protoboard. Need to triple check it though and i could improve some things i'm not totally satisfied with in component placement and ground layout too. I'll probably use a starground strategy.
I'll probably use 2134 in place of tl7x. And maybe a That 1206 in place of 275 as line input receiver, this will be determined when starting build.
About the allpass cells, well ICG's first answer to my question may well be what happen in the circuit: i've probably didn't understood R.Elliott's article and expected something more complicated than it is in reality. Anyway, i asked for help to understand this in another (great) forum and didn't have any answer.
I thought knowledge about allpass were common with experienced designers and electronicians but this doesn't seem to be the case.
I'll take some measurement to have answers once the circuit is running.
Spladski you may be right about the last series of Tannoy's pro monitors. I've seen some comments of users really disappointed about the 800. Maybe they had quality control issues, maybe this is by design but it seems their is indeed inconsistancy between S800 that is sure.
Mine are dark grey not navy blue, and the last digit is different in serial between the 2 (213103p /213102p).
I won't say i could not work on them neither that i don't enjoy listening to them, they are quite revealing in my view. Except from this thing in 1,5k ( and that i prefer closed box behavior overall) i'm quite pleased. Maybe i'm lucky and got a nice pair?
I've spent some time with 1200 too and fell in love with them. As already said liked the Active 800. Not convinced by the 600 either passive or active but it could be 'bias expectations' from my side.
I've heard and worked with Dmt 2 too. 12" and 15". Liked them both especially the 15" but they were inwall mounted and think were active filtered and biamped iirc. Not a typical use.
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.