my thoughts exactly.
May be for power tapering sort of like CBT? But then again, resistors are cheaper.
...Unless perhaps each amplifier is better...
Recently guys, I've come to love the sweet sound of Single Ended Triode amps. I'm in the midst of building a new speaker system (what can I say? my medication ran out) and I hope it will be my best yet. It uses 4 Bandor full-range drivers in a linear array - no surprise there for those that remember me - and of course no crossovers and a unique enclosure design.
Unfortunately, a single SET amp doesn't have the power that the speakers will need so I'm thinking of using a SET for each unit, giving me 4 x the power in total (in theory). I will attempt to build the SETs myself but am looking for people's views on the feasibility of this approach.
Regards
Steve
I'm still scratching my head a bit though. Granted I can see some advantages to having a dedicated amplifier per drive unit, but I can't quite see why a single amplifier should find 4 unfiltered Bandors (presumably one of the 50mm models?) an unacceptable load if it's capable of powering a single example of those units in the first place. 
Unless my calculations are out, which admittedly has been known, straightforward series-parallel wiring of two paralleled packs of two drivers in series should give efficiency gain of ~6dB over a single unit [10*log(number of parallel modules * Re]) & no impedance change, so the amplifier should in fact have an easier time than it would if it were powering just one driver. OK, so there wouldn't be any real sensitivity gain with this number of drivers & wiring configuration (10*log [Re / total array impedance]), but you wouldn't get any advantages by powering the drivers independently on that score either, so you're still back where you started, viz. no change.
Unless my calculations are out, which admittedly has been known, straightforward series-parallel wiring of two paralleled packs of two drivers in series should give efficiency gain of ~6dB over a single unit [10*log(number of parallel modules * Re]) & no impedance change, so the amplifier should in fact have an easier time than it would if it were powering just one driver. OK, so there wouldn't be any real sensitivity gain with this number of drivers & wiring configuration (10*log [Re / total array impedance]), but you wouldn't get any advantages by powering the drivers independently on that score either, so you're still back where you started, viz. no change.
Last edited:
Yes, it's 4 ohm, 50mm units. The efficiency is on the low side for a SET amp that puts out around 5 watts but gives its best for the first couple of watts.I'm still scratching my head a bit though. Granted I can see some advantages to having a dedicated amplifier per drive unit, but I can't quite see why a single amplifier should find 4 unfiltered Bandors (presumably one of the several 50mm models?) an unacceptable load if it's capable of powering an single example of those units in the first place.
In practice, I have found that I get a slight increase in efficiency by using 4 drivers, but not 6dB and mainly at the bass end.
Except that instead of 5 watts of amplifier I would have 20 watts with four of them.
Cheers
Steve
You know me well, Scottmoose!
This speaker is a 'flat panel tv matching' wall-mounted speaker*, which I'm hoping might go well with the Japanese SET lovers (of which I'm sure some are crazy enough to use two or four amplifiers). This speaker will sit far better in a Tokyo apartment than a large horn loaded.
However, I still need to convince myself that the 4 SET approach would work.
Regards
Steve
* Don't worry, it still incorporates my trademark 'inner egg', a corian front panel and, optionally, a decoupled anti-resonant transmission line which would snake along the skirting boards! (are you out there Dave, planet10?). This promises to be my craziest yet!
Last edited:
Actually, 7V's idea of having multiple amplifiers for different sections of line array might be best utilized in CBT (constant beamwith transducer) from Don Keele. He uses resistors to lower the level for higher sections, which obviously is throwing out a lot of power, as he could have each section run with amps of different gain...
I'm not surprised; practical array gain with wideband drivers is always concentrated in the LF with the top end progressively dropping off since the individual drivers can't combine into a single radiation as they move more than 1/4 wavelenth apart. Smaller the drivers the better on that score, as is the case with your Bandor systems.
That +6dB was efficiency gain rather than sensitivity, so in practice, you're not likely to get so much. Still, there's a useful reduction in the amount of work the amp has to do given that the impedance is the same, since each driver only needs 25% of the deflection of a single unit to produce the same SPL. You'd still have 5w as far as I can see if the drivers are being driven independent of each other. As far as electrical power is concerned, it's what is available to the individual drive units that counts. Total acoustical power is another matter; since you've got 4 drivers each is doing less work for a given SPL than if you just had one unit, so you're using less power from each amplifier. Pretty much exactly the same would be the case with one amplifier driving 4 series-parallel units though.
Right, that's the same use I was thinking for them (or any other variation on the power-tapering theme desired), and it's really the only significant advantage I can think of TBH, unless for some reason you needed to use combination of series-parallel drivers that would otherwise produce an unacceptable impedance load for one amplifier to handle. In both cases though, you'd need quite a bit more than four units.
Scott
Last edited:
I don't really understand your logic. One 86dB speaker driven by a, say, 4 watt amp gives out a max SPL of 92dB (at 1m). Four 86dB spakers driven by 4 x 4 watt amps should give a max SPL of 98dB. No?
I like the idea of the CBT but have grown rather fond of my 4 driver straight vertical mini-arrays and will stay with them.
Regards
Steve
I like the idea of the CBT but have grown rather fond of my 4 driver straight vertical mini-arrays and will stay with them.
Regards
Steve
OK, try to look at it this
1 drive unit = 86dB 1m/w (we'd be better using 2.83v as a baseline, but never mind).
4 drivers, each powered by its own amplifier, would give + 6dB compared to what you would get from a single unit, since we've got quadruple the cone area. That gives the 92dB, 1m/w you mention, with each of the 4 amplifiers delivering 1w into its own drive unit. And with each producing 4w into its own drive unit, we hit the 98dB mark.
So, now take 4 drivers, wired in series-parallel. That also gives 92dB 1m/w, since we still have quadruple the cone area. It's just that the single amplifier sees a more efficient speaker (6dB more efficient compared to a single driver). Thus 4w from the single amplifier will also provide a total of 98dB, or exactly the same as if you powered each of the 4 drive units with its own amplifier.
1 drive unit = 86dB 1m/w (we'd be better using 2.83v as a baseline, but never mind).
4 drivers, each powered by its own amplifier, would give + 6dB compared to what you would get from a single unit, since we've got quadruple the cone area. That gives the 92dB, 1m/w you mention, with each of the 4 amplifiers delivering 1w into its own drive unit. And with each producing 4w into its own drive unit, we hit the 98dB mark.
So, now take 4 drivers, wired in series-parallel. That also gives 92dB 1m/w, since we still have quadruple the cone area. It's just that the single amplifier sees a more efficient speaker (6dB more efficient compared to a single driver). Thus 4w from the single amplifier will also provide a total of 98dB, or exactly the same as if you powered each of the 4 drive units with its own amplifier.
Last edited:
Ok, you're right, in theory, as kr=2 is the frequency where the first concentric mode occurs - which is higher than I'd considered. In practice though, the increase in efficiency is less.
However, the wattage for each amplifier, for a given overall sound level, would be lower, and each SET would have less distortion as the load is shared. Would you agree?
Regards
Steve
However, the wattage for each amplifier, for a given overall sound level, would be lower, and each SET would have less distortion as the load is shared. Would you agree?
Regards
Steve
To an extent, yes, the extent varying with implementation. In this case, it's not likely to be vast. Ironically, I'm currently running 4 CHBW-70s, which are 84.8dB 1m/w apiece, in exactly this series parallel arrangement -the system measures at 90.8dB, although admittedly it's a 2 way, albeit with the mids covering the LF & entire telephone band out past 4KHz.
In practice you might (might) get slightly more headroom with independent amplification of each drive unit. But frankly, I rather doubt we're talking significant gains, although there aren't likely to be any acoustic downsides per se. Since you say you want to be using quality valve amplifiers, the small advantages hoped for would need to be set against substantially greater costs, both in building / purchasing and power consumption, the extra space needed for eight of said amplifiers plus presumably a couple more to drive bass units & so on & so forth. If you've got the money to burn of course, fair enough. Me, I'd probably look into making a slightly taller box if possible, & running a few more Bandors. IMO, that would bring more advantages than running individual amps per driver.
Where the independent amplification per driver concept could be interesting / come into its own would be with something like class D amps, which being compact & cheap allow such things to become a little more practical. Not what you like & not suggesting you take that route, just thinking aloud & on the hoof as it were. That would be an idea for CBT, or simpler power-tapered arrays etc. It could have some uses in ambisonics etc. too.
In practice you might (might) get slightly more headroom with independent amplification of each drive unit. But frankly, I rather doubt we're talking significant gains, although there aren't likely to be any acoustic downsides per se. Since you say you want to be using quality valve amplifiers, the small advantages hoped for would need to be set against substantially greater costs, both in building / purchasing and power consumption, the extra space needed for eight of said amplifiers plus presumably a couple more to drive bass units & so on & so forth. If you've got the money to burn of course, fair enough. Me, I'd probably look into making a slightly taller box if possible, & running a few more Bandors. IMO, that would bring more advantages than running individual amps per driver.
Where the independent amplification per driver concept could be interesting / come into its own would be with something like class D amps, which being compact & cheap allow such things to become a little more practical. Not what you like & not suggesting you take that route, just thinking aloud & on the hoof as it were. That would be an idea for CBT, or simpler power-tapered arrays etc. It could have some uses in ambisonics etc. too.
Last edited:
No way would I use SETs to drive the bass speakers.
I only use subs to fill in below about 120Hz and I wouldn't hear any improvement with SETs over a decent plate amp.I think the question for me is whether SETs would provide enough headroom with my full-range speakers before audible distortion occurs.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.