[AudioCats]

Disclaimer: HIGH VOLTAGE circuitry. Info presented is for entertainment purpose only. Do not try at home.

Starting a thread to document my approach of 336SE mod.

MassDrop is selling DarkVoice 336SE for $240.

initial impressions: (with the stock tubes and a pair of Beyer DT990-600ohm.)
-- Build quality is solid.
-- the usable volume control range is tiny. Only the first 30 degrees or so. The sound is way too loud once above that.
-- hum is obvious when music volume is low.
-- sound is clean but not a lot of micro details.

plans:
** add pigtails with 2-pin XLR connectors, connect to external box holding film output caps. (2-pin xlr = standard 3-pin XLR with one pin removed and the corresponding hole in the socket plugged, this prevents plugging into the wrong equipment by mistake)

** DC filament supply to solve the hum problem. I might do a CCS so the filament heats up slowly. This also dictates using a timer to turn on the HV only after the filament has been fully on for a while.

** increase the usable range of volume pot.

 

[AudioCats]

disassembly notes

(taking the amp apart mainly to get access to the transformer winding to take temp reading).

HV reservoir:
stock caps are three united-chemicom KMQ, 1000uf/200v (with 200 ohm resistors in between). They are mounted using steel clamps, and the clamp screws are cranked down tight. All three caps have deformed casing from uneven clamping pressure.

cap OD 30mm, there is enough room in the enclosure to accommodate up to 45mm height.
(direct upgrade: nichicon LKX 680uf/250v, https://www.partsconnexion.com/NICHICON-76192.html; or you can go a little larger and use LKX 1000uf/250v https://www.partsconnexion.com/NICHICON-76193.html )

post #6 below shows why these 3 caps should be replaced.

There are two pairs of reservoirs in the amp area, Nichicon GX, 220uf/315v. 20mm OD, 30mm L. Also clamped down, also with deformed housing.
(direct replacement: nichicon LKX 220uf/250v https://www.partsconnexion.com/NICHICON-76189.html )
(or use Mundorf MKP 8.2uf/250v, it is 22mm OD x 33mm L, installation is a little more involved but does fit ok)

 

[AudioCats]

Power transformer:
Not easy to get to it. They probably bolted on the transformer box to the chassis at the very beginning during the assembly process, then added the sub-chassis plate and other stuff.

Transformer box is glued shut using some kind of epoxy. Removing the epoxy took a bit of time.

Torroid transformer, about 85mm OD x 60mm high.

Markings:
336R-B HDB-110Tx1
Primary: 110v 50Hz red
Secondary : 135v grey 6.3v blue
shield: yellow green
NO: 210812
www.Chongzhouhengda.com 028-82250306

the Chongzhouhengda site doesn't show anything matching the dimension of this transformer.

Space inside the steel cover is 110 x 110 x 80mm.

(Hammond 263x: 115VA; 180-0-180v/250mA, 5v/3A; 2.5"x 3.68"x 3.1" high)
(Cisco CP-PWR-Cube-3: 2"x4"x1")


added thermal-couple to the inside of transformer. (Tip is tucked under the black wrapping film).

thermal-couple reading: 60C after running the amp overnight (room temp 18C).

power consumption: 66w/80VA (measured using a Kill-a-watt P3

 

[AudioCats]

Pigtail for external cap box

removed all output capacitors from the PCB.
removed output RCA's.
Added strain relief to one of the RCA holes, and a 4-wire cable.

Currently testing output caps using terminal blocks.

 

[AudioCats]

Output capacitor comparisons
factory stock tubes;
Phones are Beyer DT990-600ohm
volume control has been converted to shunt type, input resistors are Caddock TF020 30k
wiring from RCA input jacks to the pot is now 24awg silver/teflon.


Inexpensive caps:

** stock caps, 10uf/250v metalized film. Three in parallel = 30uf.
-- emphasis in mids/voice
-- bass is a bit boomy, not very deep.

** ASC metalized polyprop, 20uf/400v. I will use three in parallel to make 60uf.
( Allelectronics, $2 each, https://www.allelectronics.com/item/mkp-204/20.0-uf-400-vdc-polypropylene-capacitor/1.html )
-- strong bass with very clean texture
-- recessed mids/voice (but maybe that is normal for DT990)
-- highs are sharp, not for long listening.

** ASC LTI 4-5 metalized polyprop, 60uf/500v
(Partsconnexion, $4.5 each which is supposedly a "91% off super sale price" https://www.partsconnexion.com/ASC-81958.html )
-- strong bass but not as clean as the ASC 20uf above
-- a little warmer, better mids than ASC 20uf
-- highs are not overly sharp

so far the LTI 4-5 is the most enjoyable.

update, 11-26-2021:
after about 100hr of burn-in, the ASC 20uf/400v now wins, the sharpness has toned down a bit, mids improved; ASC LTI didn't change much, now sounds kind of dull in comparison; Not a lot of change to stock cap ( I bought the amp used, the stock caps were probably already burnt-in).

Now I just need to add a little warmth.

 

[AudioCats]

various voltage readings:

HV reservoir capacitors
#1: 203v DC (voltage rating of cap is 200v)
#2: 169v DC
#3: 135v DC
(200 ohm resistors in between the caps)

reservoir cap voltage, for input tube: 135v DC

Filament supply: 7v AC

across input tube plate resistor (30k): 66v DC
across input tube cathode resistor (1k): 2.2v DC

across output capacitors: 78-80v DC


If no tube is plugged in and the power switch is on: all HV reservoir caps see 220v DC (power input 120v AC).
I think the three large reservoir caps (200v rated) should be changed to at least 250v rated parts.

 

[AudioCats]

DC filament supply

added a 35A bridge rectifier and three 16v Panasonic FM, 2700+2700+3900uf = 9300uf.
(I tried just 2700uf x2 first and there was obvious hum, so ~9000uf is probably the minimum
(update: with my T1 I could still hear hum; added another 3900uf/16v ==> no hum)
7.1v DC at the caps.
Paralleled two 0.47 ohm/2w resistors ( equ. 0.24 ohm/4w) and put between the cap+ and filament line ==> filament voltage 6.3v DC.

power consumption: 71W (was 66w when using stock AC filament supply). up 7.5%, not that big of a deal.

The main concern is the bridge rectifier gets ho, probably should be mounted to the side wall.

on the other hand....the 60C transformer winding temp (at 66w consumption) is higher than what I am comfortable with. I will probably do an external filament supply anyway.

 

[AudioCats]

The volume Control

it turns out this one has a linear pot, no wonder the usable range is so tiny.

It is kind of a strange pot, has 4 pins: #2=wiper, #3=mid-point (47k from either end at all times).
Marking on the blue plastic part: "B104" (linear 100k)

I don't know what to say. it looks to be well made, turns super smooth, feels really good... but not made for audio use. I will probably install a basic Alps blue RK27 audio-taper to replace it.

Which means the knob will need to be changed as well (Alps has round shaft, the stock knob only fits knurled shaft).

update 11/23/2021:
found an Alps blue 50k in the parts drawer, the shaft is too long so cut it down to about 1/2". The index pin is off by 90 degrees so I cut it off (will use one of the ground wire to serve as anti-rotation).
Barely fits into place. Very tight spot. But installs ok.
Added a generic aluminum knob.

(Alps blue footprint doesn't match the little pcb. I will wire it point to point.)

 

[AudioCats]

Potting the transformer (top portion)

-- this is mostly to allow heat transfer from transformer top plate to the box shell ==> trying to reduce the transformer winding temp.

Sacrificed a BIC blue highlighter pen in the process: the pen body served as a drill stop; the end cap will plug the fill hole afterwards.

Drilled fill hole using 9/32 bit;
(Note: transformer box is up-side-down, the potting is to fill the top portion of transformer box).

mixed 250ml of EP1121 epoxy
used a large syringe (Walmart) to transfer mixed potting into the transformer box

the result..... adding potting did not help lowering the winding temp. The thermal-couple reading is now 10C higher than before. Now I really need to use an external filament supply so the transformer can run cool.
on the bright side, the "turn-on buzz" from the transformer is now gone.

 

[AudioCats]

Power resistor area

-- Since changing the cathode resistor can noticeably change the sound, I will try different power resistors as well.

there are four mounting posts in the sub-chassis; about 7mm height.
Sub-chassis has 4 square windows, each 10mmx10mm.
Distances from sub-chassis to the enclosure inner surface (sheet steel) 8.5mm;

Adding a 10mm x10mm x16mm height aluminum block will allow heat transfer to the amp enclosure (then to the thick aluminum top plate) and improve heat dissipation.
Might not be necessary, depending on the type of power resistors used.

update 11/28/2021:
test result shows there is no need to upgrade the power resistor. There was no obvious sound improvement when changed from cheap stock resistor to fancy Caddock foil type. Stock resistors are now put back to the original spots.

 

[AudioCats]

the masssDrop price seems to be depressing the price of used 336se. by quite a bit.

and some "broken/parts/repair" units are showing up on ebay.


Output capacitor test, compact options (compact enough to fit into the amp enclosure)
** due to the very limited space available, the total capacitance can only be ~30uF per channel, making it only suitable for high impedance phones (600 ohm or higher ==> Beyer T1, and 600 ohm version of DT990, 880, 770; some AKG like my k240M)

stock capacitors are 10uf/250v x3 = 30uf per channel

The most straight forward option is Auri XO 33uf/160v, 34mm OD x 46mmL (actual measurements from the two I bought). https://www.partsconnexion.com/AURIXO-79175.html

less expensive possibilities: blue Taitsu 4uf/250v metalized polyprop https://www.allelectronics.com/item/rpc-405s/4-uf-250v-polypropylene-capacitor/1.html
put 3x3 on a little perf board to make 36uF per board,
or 2x3 then add a fancier bypass cap.
for bypass cap: Amtrans AMCG 4.7uf/250v https://www.partsconnexion.com/AMTRANS-88230.html
AMCG dimensions https://www.amtrans.co.jp/archives/5904

so there are four options so far:
a) Auri XO 33uf/160v
b) Taisu blue, 3x3 (36uf/ch)
c) Taisu blue, 2x3 // AMCG 4.7uf (28.7uf/ch)
d) replace one stock cap with AMCG 4.7uf (24.7uf/ch)

listening test
-- gain tube = shuguang treasure CV181
-- output tube = RCA 6as7g
-- phone = DIY ( Beyer T1 drivers in MB Quard housing. This is the most resolving dynamic phone I have)

A) Auri XO: obviously the best, fullest sound, reveals the emotions the best. Bass a little thick, not very clean.

B) Taitsu blue, 4uf x 9 (36uf per channel)
-- sharper harder sound,
-- narrower stage,
-- cleaner bass; quantity about the same as A
-- voice not very full, every thing is clean and clear but not emotionally involving (too sterile)

C) Taisu blue 4uf x6 // AMCG
-- sweeter voice, emotionally involving now.
-- more closed-in comparing to B
-- better bass texture than B

stock caps, 10uf x3
-- very layback, not very focused. comparing to B
-- thicker bass, heavier impact comparing to B
-- not as transparent, not as live, not very exciting/engaging. comparing to B
-- softer sound, emotionally not as sterile as B.

D) AMCG // two stock caps
-- sharper sound, more upfront than stock
-- much more engaging and live comparing to stock
-- bass not as thick, cleaner

I'd say C and D are at about the same level. The phones I am using are a bit bright so maybe C is more truthful and suitable for general-purpose use.
on the other hand option-D is more practical, all you need are two AMCG 4.7uf.

But overall, Auri OX wins. (if somebody is using T1 level phones, he/she should be able to easily afford the Auri XO)

 

[AudioCats]

Power switch replacement

I bought a "power switch long lever broken off" unit off ebay.

Removing the power switch involves quite a few steps.

a) remove all front plate hex screws.
b) remove bottom plate
c) remove the four screws holding the output capacitor PCB.
d) remove all top wires connecting to the output capacitor PCB
e) desolder the transformer wires (wire coming up from under the PCB; desoldering requires using a solder sucker)
f) remove output capacitor PCB and put aside
-- this is a good time to upgrade output capacitors
g) this allows access to the two sub-chassis screws. remove screws.
h) remove volume pot knob
i) remove volume pot mounting nut
j) remove the two ground wires connecting to the volume pot PCB
k) slide volume pot back and out, to gain access to the sub-chassis screw below.
--- this is a good time to change the volume pot to a proper audio-tapered pot. Alps blue RK27 works
l) remove the remaining sub-chassis screw (power switch area)
m) slide the sub-chassis back and use a looooong screw driver (to gain a somewhat proper drive angle) to remove the four screws holding the faceplate (from the inside)
-- screw head can deform easily, try not to slip.
n) now the faceplate comes off. Remove the power switch nut and pull it out.

 

[AudioCats]

End result of thsi DV336 #2 (the "switch broken off, for repair" one)

front to back:
a) volume pot changed to Alps RK27 50K audio taper, wired as "shunt-pot", ( input resistors are in the back at the RCA jacks)
b) power switch replaced to a common short lever type.
c) signal wiring (from RCA to pot) are now 24awg silver/teflon.
d) input tube plate resistors are now 30K Caddock TF020 (ceramic chips)
e) input tube cathode resistors are now 1k Caddock TF020.
f) front power caps are now 10uf/250v Mundorf MKP (24mm OD x 33mmL, barely fits; 8.2uf ones would have been a better/easier fit)
g) small rectifier bridge and 2700uf/16v Panasonic FM as DC filament supply for the input tube. added a 1 ohm resistor in series to get the proper 6.4V. No detectable hum when using my Beyer 990/600 ohm.
h) 0.5 ohm/5w power resistor in the filament supply line (part of the 2-step filament power up circuit). the purpose of this resistor is to reduce the cold-filament current so the filaments get less of a thermal shock at power up.
i) Output coupling capacitors are now Auri XO 33uf/160v (still only really suitable for 600 ohm phones)
j) main HV power reservoirs are now Nichicon KX 680uf/250v.
k) 555 timer board to do the ~30 second delay ( 2-step filament power up).
m) two sets of input resistors at the RCA jacks, to offer different flavors: 30k Caddock TF020 (cooler) and 30k Vishay VAR Z-foil (warmer). The differences are obvious (even with Beyer 990/600).

 

[Jkas1984]

End result of thsi DV336 #2 (the "switch broken off, for repair" one)

front to back:
a) volume pot changed to Alps RK27 50K audio taper, wired as "shunt-pot", ( input resistors are in the back at the RCA jacks)
b) power switch replaced to a common short lever type.
c) signal wiring (from RCA to pot) are now 24awg silver/teflon.
d) input tube plate resistors are now 30K Caddock TF020 (ceramic chips)
e) input tube cathode resistors are now 1k Caddock TF020.
f) front power caps are now 10uf/250v Mundorf MKP (24mm OD x 33mmL, barely fits; 8.2uf ones would have been a better/easier fit)
g) small rectifier bridge and 2700uf/16v Panasonic FM as DC filament supply for the input tube. added a 1 ohm resistor in series to get the proper 6.4V. No detectable hum when using my Beyer 990/600 ohm.
h) 0.5 ohm/5w power resistor in the filament supply line (part of the 2-step filament power up circuit). the purpose of this resistor is to reduce the cold-filament current so the filaments get less of a thermal shock at power up.
i) Output coupling capacitors are now Auri XO 33uf/160v (still only really suitable for 600 ohm phones)
j) main HV power reservoirs are now Nichicon KX 680uf/250v.
k) 555 timer board to do the ~30 second delay ( 2-step filament power up).
m) two sets of input resistors at the RCA jacks, to offer different flavors: 30k Caddock TF020 (cooler) and 30k Vishay VAR Z-foil (warmer). The differences are obvious (even with Beyer 990/600).

Well done!

A lot of good info there

 

[unuselessness]

I need a new 6.3mm jack for mine, it doesn’t play audio on the right side unless kept at held at a specific angle. What replacement part do I need to order? Any pointers or input is welcomed and appreciated.