Looking at the Bulgarian schematic in the first post:
L1: frequency compensation inductor, no large current handling required, part of the feedback network so good linearity needed (preferably no iron or ferrite core), in series with a 7.5 ohm +/- 1 % resistor so presumably DC resistance must be small compared to 1 % of 7.5 ohm equals 0.075 ohm.
L2: no need for good linearity or tight tolerance, RMS current is sqrt(2) times smaller than for the output inductor because the lower part of the class C current dumper stage only conducts the current half the time
L3: I can't find L3 in this schematic
L4: output inductor, discussed above
L1: frequency compensation inductor, no large current handling required, part of the feedback network so good linearity needed (preferably no iron or ferrite core), in series with a 7.5 ohm +/- 1 % resistor so presumably DC resistance must be small compared to 1 % of 7.5 ohm equals 0.075 ohm.
L2: no need for good linearity or tight tolerance, RMS current is sqrt(2) times smaller than for the output inductor because the lower part of the class C current dumper stage only conducts the current half the time
L3: I can't find L3 in this schematic
L4: output inductor, discussed above
Yes, home made is probably best here. Would an epcos 4A be any good for L2/3 or better to stick with home made? (L3 is only in the original quad design L2 & 3 are 2uH in series in the same position as L2 in the bulgarian schematic)
With this wire i believe it would be 19 turns around 10mm for L4. (gives 1.52uH, although the wire is only rated to 1.21A. Even the 16SWG wire is only 2.74A, i'm guessing that the current rating is max sustained current)
For L1 I was looking at this as it fits the pcb perfectly with 5mm lead spacing. (but having seen the above, and it having a ferrite core, damn this is hard work home made won't fit particularly easily as it really needs to be radial.)
With this wire i believe it would be 19 turns around 10mm for L4. (gives 1.52uH, although the wire is only rated to 1.21A. Even the 16SWG wire is only 2.74A, i'm guessing that the current rating is max sustained current)
For L1 I was looking at this as it fits the pcb perfectly with 5mm lead spacing. (but having seen the above, and it having a ferrite core, damn this is hard work home made won't fit particularly easily as it really needs to be radial.)
Last edited:
Correction after thinking a bit more about the Bulgarian schematic in the first post, particularly about L1:
L1: closed-loop bandwidth limiting filter inductor, no large current handling required, part of the feedback network so good linearity needed (preferably no iron or ferrite core). DC resistance of, say, half an ohm will only change the closed-loop gain by about half a dB, no effect on the bridge balance.
L2: no need for good linearity or tight tolerance, RMS current is sqrt(2) times smaller than for the output inductor because the lower part of the class C current dumper stage only conducts the current half the time
L3: I can't find L3 in this schematic
L4: output inductor, discussed before
L1: closed-loop bandwidth limiting filter inductor, no large current handling required, part of the feedback network so good linearity needed (preferably no iron or ferrite core). DC resistance of, say, half an ohm will only change the closed-loop gain by about half a dB, no effect on the bridge balance.
L2: no need for good linearity or tight tolerance, RMS current is sqrt(2) times smaller than for the output inductor because the lower part of the class C current dumper stage only conducts the current half the time
L3: I can't find L3 in this schematic
L4: output inductor, discussed before
I expect that the Epcos will work fine as L2 and L3, but I can't be 100 % sure as there is no peak current spec anywhere. I don't know if the core could saturate during the current peaks, although that seems rather unlikely to me because of the open magnetic circuit. Maybe someone with more experience with high-current inductors can answer this?
The EPCOS inductors are using carbonyl iron cores. Core saturtaion is not a problem - rather it is winding and core losses which will limit performance. Without knowing what particular mix of core material used, nothing more specific can be stated with any confidence.
I think they use that range of epcos inductors on dadaelectronics.eu for their mods on quad amplifiers so i think it should be ok. I think the 19 SWG enamelled cable i linked to above should be ok for making L4. (19 turns around a 10mm former)
Last edited:
More than 30 years ago I built some Quad 405 clones, I have had a look at the boards and L4 is wound from what looks like about 20SWG wire 19 turns about 8mm diameter. At the time test equipment to measure distortion to 0.0001% was not readily available but I had read an article about using the amplifier under test as the active element in a oscillator, as the other components are all passive any distortion should be a measure of the amplifier distortion and using a passive filter it should be possible to remove the fundamental and measure the distortion products. I can not remember all the details but I do remember that the measured distortion was about 0.0018% so the inductor value must be quite close, I did try trimming the bridge capacitor but with little improvement.
Stuart
Stuart
I think that whilst it is desirable that the bridge circuit balances, the effects of inbalance (within the construction and component tolerance) is not truly great, this article
http://quad405.com/lips.pdf shows
@10Vrms
1.3KHz Best balance =.0042 thd with z4 shorted .0079 thd
13KHz Best balance =.012 thd with z4 shorted .032 thd
(Z4 refers to the original 405 and not the 606)
So even with 10 or even 20% tolerance components the measured thd is still going to be very low, but maybe discernible in real live listening tests.
So my take will be to use commercially available inductors, albeit for L4 I'll use 2 x 20% 3uH in parallel, thereby doubling the current capacity, and statistically likely to closer to 1.5uH than a single L4
This one looks useful EPCOS|B82111BC16|INDUCTOR, AXIAL, 3.0UH | Farnell United Kingdom
http://quad405.com/lips.pdf shows
@10Vrms
1.3KHz Best balance =.0042 thd with z4 shorted .0079 thd
13KHz Best balance =.012 thd with z4 shorted .032 thd
(Z4 refers to the original 405 and not the 606)
So even with 10 or even 20% tolerance components the measured thd is still going to be very low, but maybe discernible in real live listening tests.
So my take will be to use commercially available inductors, albeit for L4 I'll use 2 x 20% 3uH in parallel, thereby doubling the current capacity, and statistically likely to closer to 1.5uH than a single L4
This one looks useful EPCOS|B82111BC16|INDUCTOR, AXIAL, 3.0UH | Farnell United Kingdom
Yes coupling might be an issue, but then I'd modify the PCB layout to suit.
I must admit, since I am likely to be building 9 (or more) of these amplifiers I might invest in some test equipment to measure the thd, and also perhaps inductor values. This software is remarkably cheap, and can do both using the onboard soundcard form your pc, although obviously better results can be obtained with a dedicated PCI-e based card -http://www.virtins.com/page2.shtml#Instrument
So for only $39.95 you can measure frequency response, thd and much more..
I must admit, since I am likely to be building 9 (or more) of these amplifiers I might invest in some test equipment to measure the thd, and also perhaps inductor values. This software is remarkably cheap, and can do both using the onboard soundcard form your pc, although obviously better results can be obtained with a dedicated PCI-e based card -http://www.virtins.com/page2.shtml#Instrument
So for only $39.95 you can measure frequency response, thd and much more..
Ok, I've summarised my parts list now. Let me know if you spot any mistakes.
Resistors (Specials)
R7 - Vishay MRS25 metal alloy 10R 0.6W 1%
R39/40 - Vishay MRS25 metal alloy 2K2 0.6W 1%
R12 - Vishay MRS25 metal alloy 3K3 0.6W 1%
R3 - Vishay MRS25 metal alloy 5K6 0.6W 1%
R35 - Vishay PR02 metal film 12K 2W 5%
R24 - TE Connectivity ER74 WW 560R 3W 5%
R16/17 - Welwyn W21 WW 560R 2.5W 5%
R20/21/22/26/27/28/31/32/34 - Welwyn W21 WW R1 2.5W 5%
R18 - Welwyn W21 WW R12 2.5W 5%
Capacitors
Electrolytic
C7 - Panasonic FC 47uF 63V
C9/C11 - Rubycon ZL 220uF 100V
Film
C18/19 - Panasonic ECQUA X2 470nF 275V 10%
C2 - Epcos MKT B32521 330nF 100V 5%
C20 - Epcos MKT B32922 330nF 305Vac 10%
C10 - Vishay MKT372 47nF 400V 5%
C3 - Epcos MKT B32521 1uF 100V 5%
C16 - Wima FKP2 4.7nF 63V 2.5%
Mica
C1/4/6 - RS 495-802 330pF 500V 1%
C14 - RS 495-925 4.7nF 500V 1%
C8 - RS 495-701 47pF 500V 1%
Diodes
D3 - Fairchild 1N4148XTR DO-35
D4/5/6/7/8/9 - Vishay 1N4003-E3 DO-204AL
D1/2/12 - OnSemi 5W Zener V Reg 6.8V, 1N5342BG
CR2 - Semitec E-501
Inductors
L1 - MULTICOMP - AS809535-200U 20uH 35%
L2/3 - Epcos B82132A5402M 2uH 4A 0.02ohms 20%
L4 - Propower ECW1.0 enamelled copper wire 19SWG - 19 turns around a 10mm former.
Transistors
T1 - BC560CG OnSemi
T2 - MPSA43 Fairchild
T3 - BC237BG OnSemi
T4/5/6 - MPSA92G OnSemi
T7/8 - BD242C Fairchild
T9/10/11/12/13/14 - MJ15003G OnSemi
T15 - ZTX653STZ DiodesZetex
T16 - ZTX753 DiodesZetex
IC's / Other semiconductors
IC1 - TLC271CP TexasInstruments
CR1 - LM334Z National
PSU
2x Vishay Rectifier, GBPC3504A 35A 400V
12x BHC ALS30 6800uF 63V
12x BHC ETU 35mm Clamps
2x Airlink 230V to 40Vx2 300/500VA toroidal <-Still not sure if 500VA is overkill yet.
Resistors (Specials)
R7 - Vishay MRS25 metal alloy 10R 0.6W 1%
R39/40 - Vishay MRS25 metal alloy 2K2 0.6W 1%
R12 - Vishay MRS25 metal alloy 3K3 0.6W 1%
R3 - Vishay MRS25 metal alloy 5K6 0.6W 1%
R35 - Vishay PR02 metal film 12K 2W 5%
R24 - TE Connectivity ER74 WW 560R 3W 5%
R16/17 - Welwyn W21 WW 560R 2.5W 5%
R20/21/22/26/27/28/31/32/34 - Welwyn W21 WW R1 2.5W 5%
R18 - Welwyn W21 WW R12 2.5W 5%
Capacitors
Electrolytic
C7 - Panasonic FC 47uF 63V
C9/C11 - Rubycon ZL 220uF 100V
Film
C18/19 - Panasonic ECQUA X2 470nF 275V 10%
C2 - Epcos MKT B32521 330nF 100V 5%
C20 - Epcos MKT B32922 330nF 305Vac 10%
C10 - Vishay MKT372 47nF 400V 5%
C3 - Epcos MKT B32521 1uF 100V 5%
C16 - Wima FKP2 4.7nF 63V 2.5%
Mica
C1/4/6 - RS 495-802 330pF 500V 1%
C14 - RS 495-925 4.7nF 500V 1%
C8 - RS 495-701 47pF 500V 1%
Diodes
D3 - Fairchild 1N4148XTR DO-35
D4/5/6/7/8/9 - Vishay 1N4003-E3 DO-204AL
D1/2/12 - OnSemi 5W Zener V Reg 6.8V, 1N5342BG
CR2 - Semitec E-501
Inductors
L1 - MULTICOMP - AS809535-200U 20uH 35%
L2/3 - Epcos B82132A5402M 2uH 4A 0.02ohms 20%
L4 - Propower ECW1.0 enamelled copper wire 19SWG - 19 turns around a 10mm former.
Transistors
T1 - BC560CG OnSemi
T2 - MPSA43 Fairchild
T3 - BC237BG OnSemi
T4/5/6 - MPSA92G OnSemi
T7/8 - BD242C Fairchild
T9/10/11/12/13/14 - MJ15003G OnSemi
T15 - ZTX653STZ DiodesZetex
T16 - ZTX753 DiodesZetex
IC's / Other semiconductors
IC1 - TLC271CP TexasInstruments
CR1 - LM334Z National
PSU
2x Vishay Rectifier, GBPC3504A 35A 400V
12x BHC ALS30 6800uF 63V
12x BHC ETU 35mm Clamps
2x Airlink 230V to 40Vx2 300/500VA toroidal <-Still not sure if 500VA is overkill yet.
I think that whilst it is desirable that the bridge circuit balances, the effects of inbalance (within the construction and component tolerance) is not truly great, this article
http://quad405.com/lips.pdf shows
@10Vrms
1.3KHz Best balance =.0042 thd with z4 shorted .0079 thd
13KHz Best balance =.012 thd with z4 shorted .032 thd
(Z4 refers to the original 405 and not the 606)
So even with 10 or even 20% tolerance components the measured thd is still going to be very low, but maybe discernible in real live listening tests.
So my take will be to use commercially available inductors, albeit for L4 I'll use 2 x 20% 3uH in parallel, thereby doubling the current capacity, and statistically likely to closer to 1.5uH than a single L4
This one looks useful EPCOS|B82111BC16|INDUCTOR, AXIAL, 3.0UH | Farnell United Kingdom
You are also going to get some distortion from the ferrite core, no idea how much.
I note that a clone of the original 606 boards without the ground plane appear to be available on eBay:
eBay - The UK's Online Marketplace
eBay - The UK's Online Marketplace
If the boards were double sided with the ground plane i'd have considered them but since they are single sided I may as well go through the process of making my own. (for the fun of it and the experience)
I'll have a look at the pcad file though Taj, it'll do no harm to look at the options available.
As said earlier in the thread though, progress will be quite slow on this as I'll need to wait until i can afford the parts, which will be sorted in batches.
I'll have a look at the pcad file though Taj, it'll do no harm to look at the options available.
As said earlier in the thread though, progress will be quite slow on this as I'll need to wait until i can afford the parts, which will be sorted in batches.
Kei,
I've ordered the parts due to arrive today minus a couple of components that are out of stock, so can at least mod my pcb layout to suit before etching and drilling.
I've also got hold of a Pico scope so can 'fine tune' the epcos chokes for the correct tolerance using a Maxwell Bridge, and measure the distortion performance as a whole.
If all goes well I don't mind trimming some for you if that would help
I managed to get hold of about a 1m length of industrial gauge aluminium U section from a scrappy for £10, so heatsinks are more or less solved - will have to measure the thermal resistance in situ, but they 'look' about right (it's nearly 8mm thick, so T03 pins only just long enough to clear the sink+insulator+pcb!),
Hope your build goes well
Ed
I've ordered the parts due to arrive today minus a couple of components that are out of stock, so can at least mod my pcb layout to suit before etching and drilling.
I've also got hold of a Pico scope so can 'fine tune' the epcos chokes for the correct tolerance using a Maxwell Bridge, and measure the distortion performance as a whole.
If all goes well I don't mind trimming some for you if that would help
I managed to get hold of about a 1m length of industrial gauge aluminium U section from a scrappy for £10, so heatsinks are more or less solved - will have to measure the thermal resistance in situ, but they 'look' about right (it's nearly 8mm thick, so T03 pins only just long enough to clear the sink+insulator+pcb!),
Hope your build goes well
Ed
Okay 1st build done - a few issues
1. Mains round here is nearer 244 so 40v secondary nearer 43V and after the bridge and offset rail, one is pretty close to 63V == same rating as the smoothing caps... So I put 6 diodes in series to the bridge just to bring it back a bit
2. Amp is a bit noisy - can hear faint hissing in tweeter even with input shorted - so am going to replace a few of the carbon resistors with metal film, and maybe also reduce the overall gain back down to .75V instead of 0.5V
My scope is playing up so can't resolve where it is originating from at the moment
3. Toroidal transformer hum is so loud I can hear it when I enter the room :-(
Maybe there is a small offset on the mains (a DC component), so will use a bridge and couple of caps to try to solve.
4. I stuffed up the PCB layout re-orienting one half of the output stage, so a few wire links placed trackside to resolve!
Apart from that it sounds great and I think I may have overdone the heatsinks - it barely gets warm (it also weighs 11.5Kg!).
Here are some pics -
Output protection is just a fuse... will build a solid state switch for this.
Not the neatest layout, will try harder next time
View from above
Assembled - it's only going to sit behind a subwoofer bin so won't in fact be seen, but hopefully heard!
Ordered some more output transistors for the next build
Cheers
Ed
1. Mains round here is nearer 244 so 40v secondary nearer 43V and after the bridge and offset rail, one is pretty close to 63V == same rating as the smoothing caps... So I put 6 diodes in series to the bridge just to bring it back a bit
2. Amp is a bit noisy - can hear faint hissing in tweeter even with input shorted - so am going to replace a few of the carbon resistors with metal film, and maybe also reduce the overall gain back down to .75V instead of 0.5V
My scope is playing up so can't resolve where it is originating from at the moment
3. Toroidal transformer hum is so loud I can hear it when I enter the room :-(
Maybe there is a small offset on the mains (a DC component), so will use a bridge and couple of caps to try to solve.
4. I stuffed up the PCB layout re-orienting one half of the output stage, so a few wire links placed trackside to resolve!
Apart from that it sounds great and I think I may have overdone the heatsinks - it barely gets warm (it also weighs 11.5Kg!).
Here are some pics -
Output protection is just a fuse... will build a solid state switch for this.
Not the neatest layout, will try harder next time
View from above
Assembled - it's only going to sit behind a subwoofer bin so won't in fact be seen, but hopefully heard!
Ordered some more output transistors for the next build
Cheers
Ed