The parts kit has been our of stock for ages. Has anyone compiled an upgrade part list to order direct from Digikey? Would like to try some fancier parts and see if there is an improvement.
Mostly caps but would use high precision resistors if extreme accuracy was helpful. Did this with the ACA Amp and thought it might have helped. Any other parts you can suggest? Don't mind spending a bit. My two stereo systems are based on this linestsge with dual PSUs and they are so powerful they make me want to listen.
Yes, could you use “better parts” to see if there’s a difference? Sure!
The real issue is defining what “better” would be in this application. The signal path of BA2018 is DC coupled, and that means that the only capacitors in the entire circuit are local power supply decoupling, and a cap or two in the feedback circuit to keep oscillation controlled (I.E., not happening) at RF.
This means that the audio path is all resistors and transistors. So what will you change? Transistors could be matched, though that’s not necessary, perhaps getting a bunch of the transistors and matching them all could make a difference. That would be a fun exercise regardless. As for the resistors, the kit already comes with 1%, low-noise, precision metal film, so anything else would be chasing 0.1% resistors, or buying something boutique and weird. Not sure either would make much difference.
PSU would be where I’d spend effort and attention, even though the circuit sounds great with someting simple like a 317/337 set. SuperReg, SilentSwitcher, a shunt like Salas BiB, one of @rhthatcher PSUs would all be worth looking into if you wanted something higher performance.
As for availability, restock is imminent…wait a week for the kits to hit the store.
The real issue is defining what “better” would be in this application. The signal path of BA2018 is DC coupled, and that means that the only capacitors in the entire circuit are local power supply decoupling, and a cap or two in the feedback circuit to keep oscillation controlled (I.E., not happening) at RF.
This means that the audio path is all resistors and transistors. So what will you change? Transistors could be matched, though that’s not necessary, perhaps getting a bunch of the transistors and matching them all could make a difference. That would be a fun exercise regardless. As for the resistors, the kit already comes with 1%, low-noise, precision metal film, so anything else would be chasing 0.1% resistors, or buying something boutique and weird. Not sure either would make much difference.
PSU would be where I’d spend effort and attention, even though the circuit sounds great with someting simple like a 317/337 set. SuperReg, SilentSwitcher, a shunt like Salas BiB, one of @rhthatcher PSUs would all be worth looking into if you wanted something higher performance.
As for availability, restock is imminent…wait a week for the kits to hit the store.
Last edited:
you might source Dale resistors if they don't come in the new kits but really and truly the units in the kits are pretty good.
there are no capacitors to speak of in the circuit path.
money spent on regulated power supplies for each channel -- super regs or silent switchers -- would be worth your time and money. I'm planning on super regulators.
and attenuators...
the only substitution I'm making is a side@-actuated trim pot because I'm mounting the boards vertically side by side.
there are no capacitors to speak of in the circuit path.
money spent on regulated power supplies for each channel -- super regs or silent switchers -- would be worth your time and money. I'm planning on super regulators.
and attenuators...
the only substitution I'm making is a side@-actuated trim pot because I'm mounting the boards vertically side by side.
@6L6: Am already using dual linear PSUs (Randy's design) and they make a significant improvement. Your explanation of the circuitry design is very helpful.
@mhenschel: Randy's PSUs are nicely regulated and the flex version I have been using for years can be trimmed from 15 to 24 vdc. Use 18 vdc for the line stage as this was recommended by Wayne.
The main issue for dual PSUs is transferring the 6 lines from the outboard box to the preamp one; for this I use amphenol connectors from Mouser. They are very expensive but since I use them for all phono and preamp stages the cables can be switched as needed. The specifics for this design are given in the Pearl 3 thread.
@mhenschel: Randy's PSUs are nicely regulated and the flex version I have been using for years can be trimmed from 15 to 24 vdc. Use 18 vdc for the line stage as this was recommended by Wayne.
The main issue for dual PSUs is transferring the 6 lines from the outboard box to the preamp one; for this I use amphenol connectors from Mouser. They are very expensive but since I use them for all phono and preamp stages the cables can be switched as needed. The specifics for this design are given in the Pearl 3 thread.
I'm finally getting down to the point where I can start drilling out screw holes to mount parts for my BA2018. But I have questions regarding ground connections.
1) I believe the BA2018 and Muses board get their ground from the power supply, so I don't think I would connect either to a star ground (like the transformer shield would) inside the chassis. Correct?
2) I have Cinemag output transformers for balanced output. Should these be mounted directly to the chassis, or should they be isolated from the star ground/chassis? I think these would also be isolated from the chassis ground.
3) Bonus question: Does the ground to the balanced output connector come from the ground on the BA2018 board.
Thanks!
Sneak peek...
1) I believe the BA2018 and Muses board get their ground from the power supply, so I don't think I would connect either to a star ground (like the transformer shield would) inside the chassis. Correct?
2) I have Cinemag output transformers for balanced output. Should these be mounted directly to the chassis, or should they be isolated from the star ground/chassis? I think these would also be isolated from the chassis ground.
3) Bonus question: Does the ground to the balanced output connector come from the ground on the BA2018 board.
Thanks!
Sneak peek...
1) Transformer shields connect to, and should be considered as, chassis. Re: the starground part of that question, I don't understand.
2) See above. Connections to the jacks and BA2018 should float.
3) 'Ground on the balanced connector' meaning the lug? Or pin 1..?
2) See above. Connections to the jacks and BA2018 should float.
3) 'Ground on the balanced connector' meaning the lug? Or pin 1..?
6L6, you've led me to the answer yet again. I ran into this today:
https://electronics.stackexchange.c...of-an-xlr-connector-not-be-connected-to-pin-1
...which includes the text, "...the pinout for XLR connector is such that pin 1 is common audio ground reference pin, pins 2&3 are the balanced audio pins, and depending on how you want to count it, pin 4 is the connector shell." It goes on from there but it seems to be a good reference for what I am doing.
Thanks again!
https://electronics.stackexchange.c...of-an-xlr-connector-not-be-connected-to-pin-1
...which includes the text, "...the pinout for XLR connector is such that pin 1 is common audio ground reference pin, pins 2&3 are the balanced audio pins, and depending on how you want to count it, pin 4 is the connector shell." It goes on from there but it seems to be a good reference for what I am doing.
Thanks again!
Kits are now back in stock in the diyaudiostore
Question about resistor size and performance.
Ordered a large number of resistors from DigiKey using the 2018LS BoM part numbers. They sent, largely, very small resistors that are 1.60mm x 3.60mm (the larger size is 3.68mm x 8.74mm). These smaller resistors are 1% and 0.4W but look dangerously small. Checked on the Internet about the differences between larger and smaller resistors and found that smaller resistors (such as these) have two minuses -- they dissipate less heat and are noisier (so-called Johnson noise -- present as the result of any resistance and a partial determinant of S/N range).
So...
If you are shooting for a premium build should you use larger resistors in the Wayne 2018 linestage? The smaller ones have the advantage of taking up less PCB space.
An enquiring mind who would benefit from this forum's experience...
Ordered a large number of resistors from DigiKey using the 2018LS BoM part numbers. They sent, largely, very small resistors that are 1.60mm x 3.60mm (the larger size is 3.68mm x 8.74mm). These smaller resistors are 1% and 0.4W but look dangerously small. Checked on the Internet about the differences between larger and smaller resistors and found that smaller resistors (such as these) have two minuses -- they dissipate less heat and are noisier (so-called Johnson noise -- present as the result of any resistance and a partial determinant of S/N range).
So...
If you are shooting for a premium build should you use larger resistors in the Wayne 2018 linestage? The smaller ones have the advantage of taking up less PCB space.
An enquiring mind who would benefit from this forum's experience...
way back.in this thread this was discussed a bit. the expert advice was that these were really good resistors size notwithstanding.
on the other had, it was also presented as conventional wisdom that the larger Dale units -- those light brown turds -- were the gold standard... but harder to come by.
we've gone through a lot of recommendations over the years... Corning RN60, Resista, Holco, Vishay... and that doesn't even get us up to date with SMDs.
I expect the experts will suggest that these micros are just fine for a BA2018
on the other had, it was also presented as conventional wisdom that the larger Dale units -- those light brown turds -- were the gold standard... but harder to come by.
we've gone through a lot of recommendations over the years... Corning RN60, Resista, Holco, Vishay... and that doesn't even get us up to date with SMDs.
I expect the experts will suggest that these micros are just fine for a BA2018
A large contributor to the sound of a circuit is the design of the circuit. The brand of the resistors in my opinion has much less or very little effect on the sound of the circuit.
I buy no name made in China resistors from a local electronics store when I need resistors between Mouser or Digikey orders, and they go into my builds, and they sound good. I do not spend time worrying about the sound of resistors. 🙂
I buy no name made in China resistors from a local electronics store when I need resistors between Mouser or Digikey orders, and they go into my builds, and they sound good. I do not spend time worrying about the sound of resistors. 🙂
Those teeny resistors were fantastic, the only reason I didn’t use them a 2nd time was they were so challenging to read.
You’re welcome to worry about resistor quality if you so desire, however I’d suggest it’s not worth one iota of concern.
You’re welcome to worry about resistor quality if you so desire, however I’d suggest it’s not worth one iota of concern.
Want to build the best BA2018 for my purposes and 6L6 suggested above possible improvements from transistor matching and 0.1% resistor selection. Will report progress in three posts: matching, resistors, and audio analysis.
Here's the first on transistor matching.
Knew nothing about transistor matching before starting this project and Jim told me that the Peak SOT23 would be necessary for measuring the 2SK209 SMDs. Saw that this connected to Peak's DCA75 component analyzer so also ordered this.
The Peak analyzer is a great device and is way, way cool. It provides a ton of info on any part and connects to your computer for curve tracing. Easy to learn; cf. 'Peak Setup' below.
Ordered 100 of each of the 2SK209, KSA992F, KSC1845F and LTL-1CHE parts from DigiKey. Upon testing found variance to be huge. Decided to match eight parts for each class so that both channels were completely matched and therefore the soundstage would benefit. The KSA992Fs came loose and were the most variant – over 75 had to be measured before a set of eight could be assembled. KSC1845Fs were on cut tape and had less variance for match, about 40. And the 2SK209s came on reel tape and required 35 checks for final selection. Decided in the future to always order on cut tape for measuring/marking ease.
Use a 4-week pill organizer to organize parts into plastic bins.
The 2SK209s were matched by Idss values at 4.75mA. This would be impossible without the SOT23Test Adapter. Note in the 2SK209 JPeg below the 4 curve tracings. Each line merges 8 individual traces at 4 voltage levels. The small variances that occurs within each line are subtle and different than the PnP transistors. This Jpeg illustrates Peak's matching capabilities that are impressive.
The KSA992Fs are PNP silicon BJTs and were matched by the hFE value. All other values (e.g. VBE) follow suit. See the KSA992F Jpeg below. The curve matching is especially tight here at 505hFE.
KSC1845Fs, also PNP silicon BJTs, are also matched with hFE values (395). See the KSC1845F Jpeg below. Again, nice and tight overlapping among the seven traces for each Ic level.
The LEDs were measured just as a curiosity; it turns out their variance is low and a value of 1.921 volts covered many in the package.
Now have all of the transistors matched and ready for insertion into the PCB. Will be getting the high precision resistors this coming week and will report again once they are installed. Finally, will provide an audio comparison with the boards already being used to see what changes take place in the premium build.
Have added a Project Cost spreadsheet below for any that are interested in using some of these techniques.
Here's the first on transistor matching.
Knew nothing about transistor matching before starting this project and Jim told me that the Peak SOT23 would be necessary for measuring the 2SK209 SMDs. Saw that this connected to Peak's DCA75 component analyzer so also ordered this.
The Peak analyzer is a great device and is way, way cool. It provides a ton of info on any part and connects to your computer for curve tracing. Easy to learn; cf. 'Peak Setup' below.
Ordered 100 of each of the 2SK209, KSA992F, KSC1845F and LTL-1CHE parts from DigiKey. Upon testing found variance to be huge. Decided to match eight parts for each class so that both channels were completely matched and therefore the soundstage would benefit. The KSA992Fs came loose and were the most variant – over 75 had to be measured before a set of eight could be assembled. KSC1845Fs were on cut tape and had less variance for match, about 40. And the 2SK209s came on reel tape and required 35 checks for final selection. Decided in the future to always order on cut tape for measuring/marking ease.
Use a 4-week pill organizer to organize parts into plastic bins.
The 2SK209s were matched by Idss values at 4.75mA. This would be impossible without the SOT23Test Adapter. Note in the 2SK209 JPeg below the 4 curve tracings. Each line merges 8 individual traces at 4 voltage levels. The small variances that occurs within each line are subtle and different than the PnP transistors. This Jpeg illustrates Peak's matching capabilities that are impressive.
The KSA992Fs are PNP silicon BJTs and were matched by the hFE value. All other values (e.g. VBE) follow suit. See the KSA992F Jpeg below. The curve matching is especially tight here at 505hFE.
KSC1845Fs, also PNP silicon BJTs, are also matched with hFE values (395). See the KSC1845F Jpeg below. Again, nice and tight overlapping among the seven traces for each Ic level.
The LEDs were measured just as a curiosity; it turns out their variance is low and a value of 1.921 volts covered many in the package.
Now have all of the transistors matched and ready for insertion into the PCB. Will be getting the high precision resistors this coming week and will report again once they are installed. Finally, will provide an audio comparison with the boards already being used to see what changes take place in the premium build.
Have added a Project Cost spreadsheet below for any that are interested in using some of these techniques.
Attachments
Wow!
Wonderful dedication, I bet it’s going to be worth the work, looking forward to hearing your impressions.
🙂
Wonderful dedication, I bet it’s going to be worth the work, looking forward to hearing your impressions.
🙂
I had to buy some spares as I inadvertently soldered them in with the wrong polarity and damaged them when trying to desolder them. i should have a few spares, but can't check at the moment as am traveling until next week. If you need them urgently, i think I got mine from TME:
Transfer Multisort Elektronik Limited Coleshill House, Suite 1C, 1 Station Road Coleshill, Birmingham, B46 1HT, United Kingdom VAT No. 259 1715 85, Company Number: 10482843 Tel.: +44 (0)1675790026, e-mail: office@tme-uk.eu
| No. | Item | Ordered Qty | Conf. Qty | %VAT | Net price | Net value | Estimated time of delivery** | Notes |
| 1 | LTL-1CHE LED;3.1mm;red;5.6÷12.6mcd;60°;F ront:convex;2÷2.5V ROHS Mfr: LITEON Original part number: LTL-1CHE | 10 PCS | 10 PCS | 20 | 9,88 GBP /100 PCS | 0,99 GBP | In stock | |