Modifying the Indeed G2 hybrid headphone amplifier In essence the G(eneration)2 is similar to the older versions. There are a few differences compared to the other amps: better PCB (ground plane) relocation of some parts better quality power switch input selector switch (3.5 jack - RCA) higher class-A current in/output RCA connectors on the back higher value (20k) logarithmic volume potentiometer 6V - 12V heater selector switch so more tubes can be used small bypass caps on output caps output disable at start-up (only when the power switch is used) different output resistance (accounts for the biggest difference in SQ) Picture from sellers e-bay website This amplifier can, however, still be improved on some points. the power decoupling capacitor is rated at a too low voltage (25V) the LM317's have an undersized/rated cooling fin the frequency range is too limited (makes it sound 'warm' though) and can be expanded distortion can be decreased output power can be increased crosstalk can be eliminated temperatures in the amp can be lowered l owered 300mA/365mA tube tube rolling made better output disable function improved A couple of different mods can be done to achieve these goals. 1: Add cathode caps 470 F/10V (expands the frequency range/attack when IRL530/510 are fitted) 2: Change output resistor values (between 22 and 120 Ω) G2's value is 30.1 Ω. 3: Replace the 6,800 µF/25V for a 35V capacitor (same size will be 3,300 µF/35V) 4: Replace output caps with better quality or bigger value (lower frequency limit expands) 5: Replace IRF630 to IRL510/IRL530 (frequency range/attack and crosstalk/distortion improves) 6: Replace LM317 to higher grade LM317A (not really needed and won't improve the sound) 7: Feed heaters from own power supply (involves cutting of traces, most difficult mod but worth it !) Easy to do mods: 1 to 4 more difficult because some de-soldering skills/insight is needed: 5 to 7 More experience recommended: 7 These mods and a few others will be discussed on the following pages. Not specifically as in the order mentioned above. Pick any of them you like or.... do them all as can be seen in the end result on the right.
The large cooling fins have the MOSFETs and LM317's fitted on them and is needed for mod 7. Another option is to leave the MOSFETs on their cooling fins and place the LM317's on a cooling baseplate as is done with the Miridiy.
Schematics of the original Indeed G2 amplifier:
Because the used IRF630 MOSFET with it's high Qgs (the charge that is needed for the Source to follow the Gate) this is a rather 'poor' point of these amps (unless you prefer the rolled of highs). A lot of people find the rolled of highs the big charm of this amp, makes it sound... warmer. I prefer my amps straight within the audio band meaning 10Hz – 30kHz (-0.5dB) is required. Well below 20Hz and above 20kHz which is considered the audible bandwidth. The specifications of the G2 as specified by the seller: • In Put Power: DC24V 900mA • Input Impedance: 100KΩ • Output Impedance:12~600 Ω • Gain: 20dB • Frequency response:10Hz-60KHz +/- 0.25dB • Input Impedance: 100KΩ The reality is different: • Input Power: DC24V 580mA • Input Impedance: 22KΩ • Output Impedance: 30 Ω • Gain: 26dB • Frequency response: 8Hz-10KHz +/- 0.25dB It makes you wonder the validity of the other specifications they provide. The specified values have absolutely no relevance to the actual product.
The gain is determined by the tube's gain itself as there is no feedback in this amp. There can be differences in gain between the 2 halves of a tube (so the 2 channels) of 1dB. The drop in the higher frequency region is caused by the IRF630 and not limited by the tube. This is because the CCS (constant current source) in the anode circuit has a very low current due to the low power supply voltage (most tube amps run on voltages between 100 and 300V) This low current (400 µA) is not able to charge the GS capacitance (Qgs) of the IRF630 quick enough for the higher frequencies. This means you can get the frequency response up by replacing the IRF630 with another MOSFET. You can actually tailor the frequency range by choosing the MOSFET. If you like the -0.5 dB point to shift to 20kHz use an IRF530 If you like the -0.5 dB point to shift well above 30kHz use an IRF510. It is better to opt for IRL type MOSFETs instead of IRF. The crosstalk is caused by the LM317's that run out of their minimum required operation voltage on the lower half of an audio signal above a certain output level. This causes distortion that is NOT pleasing to the ears. (see middle picture on page 2) Preventing the voltage to drop below the minimum level can be done by increasing the anode voltage or by choosing a MOSFET type with a lower voltage drop in Source follower mode. With IRF types this is 4V and IRL types this is 2V. A 2V gain in amplitude (translates to 1.4V higher undistorted output) Added bonus.. the filament (heater) of the tube is guaranteed to get the optimal voltage. So by choosing an IR L530 (20kHz – 0.5dB) or IR L510 (above 50kHz – 0.5dB) you can increase the bandwidth AND lower the distortion when you have the anode bias-voltage set at 13 to 14 Volts. In modification 5 changing the MOSFETs is discussed more thoroughly. When replacing the MOSFETs with IRL types there will be more voltage across the LM317's and because of this reason they will become much hotter then before easily reaching 60 oC or 70oC. Another mod that expands the high frequency limit even further is the placement of cathode caps on the 2k bias pots (modification 1): The caps used in this amp are 470 µF/10V. They can be placed on the bottom of the PCB or, as in this case, on the topside because there was plenty of room available when the LM317 and MOSFETs were relocated to the outside of the amp. Cathode caps mounted next to the bias pots. The + is soldered on the top side to the bias pot. The – is soldered on the copper side by scratching away some lacquer exposing the copper (see below)
After replacing the IRF630 with IRL510 and adding 2 bypass capacitors on the cathode bias pots the frequency range has become a very respectable: 0,9Hz-100kHz (-0.5 dB) / 0.3 Hz-250kHz (-3dB)
The green band in the bottom of the graph on the right shows the width of the audible range of a young person. -0.5 dB is considered barely noticeable -3 dB is what is usually given in specifications of amplifiers / equipment.
Illustrating the increase in bandwidth works best by using 10kHz square waves. The sharper the rise/fall of the edges the better the 'attack'. Above trace (1) source = what it should look like, trace (2) below is the amps output waveform.
Before shot with IRF630 fitted. rise/fall time approx. 0,2V/ µS The edges of the square wave are rounded off illustrating the limited bandwidth. The output signal (trace 2, the lower trace) should look more like the upper trace (1) which is the source signal applied to the input of the amp.
'after' picture with IRL510 This one is at maximum output power (not yet near clipping point) ... even more impressive. No ringing or overshoot VERY nice indeed. rise/fall time approx. 4V/µS
… Just for the fun of it... a 100kHz square wave
modification 4: To expand the low frequency limit you can replace the 2 output caps (1,000 µF/25V) with larger ones. 2,200µF/25V will lower the frequency range to half that of the original frequency. This is not needed but it can't hurt. The least that can be done is replacing the output caps to 1,000 µF/25V/105oC types. These will last longer because they are near very hot components. A note concerning the parallel capacitors of 100nF that are placed in parallel to the output caps in the G2 amps. Technically 100nF are used in parallel to big decoupling capacitors in power supply lines because the inductance of the capacitor (it's made of rolled up foil) increases the ESR above say... 1 MHz. This is important for the proper operation of fast ICs and switching things (SMPS, clock, digital ICs) proper decoupling of power lines can MAKE or BREAK the audio quality of these devices. This amp has it's -3 dB point at 200kHz (fully modded, otherwise 30kHz) Basic calculations say: impedance of 100nF (0.1 F) @ 20kHz = 80 Ω impedance of 1,000 µF @ 20kHz = 0.008 Ω. This is 10,000x times smaller, well ...actually slightly less... namely the ESR value of that specific capacitor to be exact. These 'resistances' are in parallel so easy to guess where the majority of the (audio)current flows. To put it in perspective: 99.999% of the audio signal (current) passes through the 1,000 µF capacitor and 0.001% through the 100nF. in dB's we are talking about 80dB difference ! Even when a 1 µF capacitor is placed in parallel it still has 8 Ω (1,000x times higher) resistance. This same ratio is there throughout the entire audio range. At 20Hz a 1,000F cap = 8 Ω... 100nF = 80 kΩ. These ratios become smaller above 100kHz and will be in favour of the 100nF above say... 10 MHz.
Square wave of 10Hz showing the lower frequency limit to go very low 0,9Hz (-0.5 dB) / 0.3 Hz (-3dB) ! Very low indeed for a capacitor coupled tube amplifier.
Modification 3 is another important modification and consists of replacing the 6,800 µF/25V capacitor for a 35 Volt type capacitor. There is 24Volt across this capacitor and the 25V rating for it is a bit 'tight' A 6,800µF/35V capacitor will be a lot bigger in this case and won't fit. Because of a simple formula (called CU product) the physical size of the capacitor is directly related to it's value C ( µF) and the working voltage U. A mechanically equally sized capacitor therefore will be around 3,300 µF/35V. In this amp a 1,000 µF/35V capacitor is fitted which is more then enough for this purpose. Measurements showed that there is NO change in the power bandwidth even when a 220 µF/35V capacitor was used.
modification 2: This is a very simple way to change the sound of this (actually any) headphone amp. The output resistance has a rather big influence on the sound in conjunction with the used headphone. This depends on the impedance characteristic of the used headphone as well as it's nominal impedance.
To make it sound like the earlier model amps replace the 30 Ω with a 47 Ω resistor. You can also make the resistors 120 Ω which is a standard for a lot of modern headphones. The output power will be considerably lower in this case. 30.1 Ω, that is currently in there, works fine too. These are the blue resistors (one is underneath the output plug) For a 'universal' value 68 Ω seems rather optimal.
EXTRA modifications:
Here are the output resistors located.
One modification consists of removing the input selector switch. If you don't need or like this selector (switches between the 3.5mm and RCA inputs) you can take them out and make connections where it came out. In this case it was needed to make room for the cooling fins on both sides. I prefer both the RCA and 3.5mm are always connected so it was no problem. Switches have a tendency to cause problems after many years anyway. Especially the cheap type used in this amp. The connections made after the extraction can be seen on the right.
Output enable relay modification: The G2 amplifier has a relay that enables the output signal about 20 seconds after the amp has been switched on. This is to prevent a loud 'pop' heard on the older versions of this amp when switched on. The drawback of the used circuit (this way it was easy and cheap to make) is that when the amp is not switched on using the power switch at the back but rather by switching on the power supply with the power switch on the back of the amp already in the 'on' position the loud 'pop' is still present. Furthermore this circuit needs some time after the power supply has been turned on to work correctly. Also when the amp is switched off the relay doesn't disable the output immediately resulting in a short DC voltage 'pop' on the output(s). The behaviour of this circuit can be altered so it doesn't need time to work when the power supply is connected while the amp is still off AND it switches off immediately when the power switch is set in the 'off' position. 2 Components need to be changed. These are time determining parts so if you want to change the 'disable output' time.. this is your chance. Replace the upper resistor (originally 220 kΩ) to 220 Ω. This makes it switch off the outputs much faster preventing a switch-off 'pop' to occur. Replace the 100 µF capacitor for a 47 µF/35V type OR replace the resistor closest to the transistor (100 kΩ) for a 47k resistor. This is needed because the 220 kΩ resistor that was changed affects the switch-on time too. If you want the output to be enabled somewhat faster use even smaller values (minimum value 22k Ω !)
Tube rolling : This amp has a switch (6V - 12V) in order to let you 'roll' more types of tubes. Make sure this switch is in the correct position (depending on tube type) before switching it on ! The heaters in the tube (filaments) in the G2 are fed from 2 current sources made up of LM317's with the a resistor between the 'out' of this regulator and the 'adjust' pin. Across this resistor (4.3 Ω) there should always be 1.25 Volt (the reference voltage of this IC) When the audio output is driven beyond a certain level the reference voltage becomes too low and the current source stops to function as it should and gives off (unpleasant) asymmetrical distortion. Both class-A output stages where these current sources are in have 1.25/4.3Ω=290mA flowing. A part (230mA) is dissipated by the 2 pieces of 68 Ω/3.75W resistors on top of each other near the tube, the other part (350mA) is for the tube. The 6922 tubes and ECC88 (the same) have 365mA/6.3 volt heaters... a perfect match. 365mA tubes will be receiving close to the appropriate voltage (6.2V) When a 300mA tube is used the voltage across is will be 7.3V No problem with PCC tubes but for 12AT7, 12AU7, 12AX7 = ECC81 / 82 / 83 (12V setting) this is another matter as these tubes will have about 1V (20%) over-voltage on their heaters as they are 300mA tubes and this will shorten their lifespan. Furthermore because of this higher heater voltage the crossover distortion when I RF are used will already be occurring at lower output levels. Increasing the bias voltage with 1 to 2V helps. The gain of this amp varies with the tube also as would be expected because the tube's 'S' (mA/V @ Ugk = 0.5V) is the gain determining part as there is NO feedback in this amp. Varying from 16x (24dB) to 21 times (26dB) with differences between tube halves (thus L and R) up to 1dB difference. Other tube types will have different gain and thus the amp has a different gain too (volume pot setting) Be careful when tube swapping. By altering the heaters power supply from current sourced to voltage fed these problems are over. PCC tubes will be slightly underpowered though but still work and the tubes lifespan won't be shortened. To address the crosstalk (to make it go away completely instead of pushing it more back to the point where it is less obvious) and distortion even further the heaters have to be taken out of the output stage and fed directly from a 6.5V switch mode voltage regulator so this is an added bonus. The used switching voltage regulator is: TRACO POWER, TSR 1-2456 The voltage is slightly too high but a Schottky diode lowers this voltage to 6.25V. A Normal voltage regulator (linear, not switching) would get extremely hot and the amp in total would draw MUCH more power. When this mod is done the LM317's on their (too small) cooling fins would unsolder themselves (if they did not switch off at a certain temperature) so will have to be cooled MUCH better. I did this by relocating them onto the large cooling fins on both sides of the amp but it can also be done with an aluminium bottom plate below the amp (replacing the acrylic bottom plate) Added bonus: This plate is grounded and also acts as a shield against hum / interference etc.
The LM317's MUST be mounted with insulating sets as the outputs are not allowed to connect to ground (output pin of the LM317 is connected to the metal mounting part of it's housing) On the right there is an example of how this can be done with an aluminium bottom plate replacing the polycarbonate bottom plate. Notice the insulating kit (plastic ring that isolates the screw PLUS silicone sheet under the LM317A.) Using an isolation kit is of great importance.
Modification 5: De-solder the IRF630's 3 pins using de-soldering braid or a de-soldering gun. Remove the screw that holds it (it is tightened rather firmly). You will have to grip the metal bar between the 2 cooling fins firmly with a small wrench or big pliers. Pull out the IRF630. Apply a very THIN layer of silicone cooling paste (scrape it off the IRF630 if you don't have it) and fit the IRL530 or IRL510 mechanically (screw it tightly onto the cooling fin). The white stuff is hard to get off your hands and wash out of clothes so.. be careful. Now solder the pins onto the PCB and cut off the excess pin. NOTE: The cooling fins have 24V on them when this amp is active as no insulating kit is used ! An insulating kit can be used to 'cure' this shortcoming. Because I find the temperatures of the MOSFETs on the high side I decided to place big cooling fins on both sides of the amp so I took the MOSFETs off the board and mounted them on big cooling fins using insulating kits. ( important ). The cooling fins on the PCB are removed which creates a nice open space. Do this by de-soldering the heat sink pins. Place header pins where the MOSFETs (and LM317) had been mounted where the wires can be soldered on easily. The insulating kits can be seen clearly on this picture.
Modification 6: De-solder the LM317s 3 pins using de-soldering braid or a de-soldering gun. Pull out the LM317 along with it's small cooling fin. You can replace the LM317 for a higher grade LM317A. Won't make any audible differences so you might want to leave i t in place. Saves you buying this part and de-soldering. If you decide to put the LM317's on a bottom cooling plate (picture on the left) or like in this amp on the cooling fins on the side (picture above) then they will have to be taken off anyway. Take out the screw and salvage the LM317 if you are not going to mount an LM317A. Shorten the pins to a smaller length (see picture above). I changed the class-A currents to a lower value by changing the LM317 resistors to a bigger value of 5.6 Ω, this way the total power consumption remains the same, the temperature will be slightly lower and the maximum output currents can still be reached while remaining in pure class-A. Another (totally unnecessary) modification I did was replacing the blue LED with an orange one. I like the colour better then the posh blue light. You can replace this high efficiency blue 3mm LED with any colour you like.. yellow, green, orange, red or white. DO NOT replace the red LED for another colour !! This is an important part of the anode current source and if you change the colour you change this current too because of the different voltage drop across these LEDs. It has a voltage drop of 1.8V creating the reference voltage for the constant current source (CCS). I find the LEDs to glow too bright so I changed the resistor value (1k5 resistor is located under the tube socket !) with a 3k3 resistor to dim it a little. I placed this on the bottom of the PCB. I soldered the cathode of this orange LED directly to ground but you can also solder it in it's original place. The cathode of the LED is then connected to the heater which is no problem even when mod 7 is done. If you want the LED to connect with the ground plane on the PCB you will have to scratch of the soldering mask at that point (green lacquer) with the tip of a small screwdriver and expose the copper. Tin this area with solder directly after exposing the bare copper.
Modification 7: Build a power supply for the heaters using a switch-mode regulator. I chose the TRACO POWER TRS 1-2465 (6.5V output) and wide range input. It has a similar pin-out as a standard 3 pin regulator so it can be swapped out ease. A NON-switching (normal linear regulator) can NOT be used as this would get extremely hot and would need a huge cooling fin, the switch-mode regulator will stay cool... and needs no cooling fin. ( Farnell order code: 169-6321 ) For this modification a trace has to be cut near pin 4 of the tube foot and some connections to the ground plane have to be made. The 2 big 3W 68 Ω resistors between the tube and volume control must be removed and also the wire underneath these 2 resistors MUST be removed. Scratch of the soldering mask at that point (green lacquer) with the tip of a small screwdriver and expose the copper. Tin this area with solder directly after exposing the bare copper.
Cut the trace in the red circle shown above. The heaters power supply will come from the green wire that is connected to the regulator. MAKE sure the wire underneath both 68 Ω resistors is cut or removed !
Ground plane connections have to be made on these spots. NOTE: the blue LED does not need to be soldered to the ground plane and can remain in it's original position.
Mount the TSR 1-2465 with it's components as closely as possible to the Big power capacitor with as short as possible connections ( important). The Schottky diode is guided through the black rubber sleeve with the cathode connected to pin 4 of the tube and the anode connected to the output pin of the regulator. The input pin is connected via a 100 H inductor onto the + pin of the power capacitor and the middle pin (ground) is connected directly to the – pin of the power capacitor. Details about this part are on the next page. There are 2 small SMD 100nF capacitor placed between these pins (the small brown parts). Between the input and ground a 10 F/50V capacitor is connected. This, together with the inductor and short connections make sure switch-mode spurious signals can not contaminate the signal paths which it will if you mount the regulator in a different position. Between the out of the regulator and the ground pin a 470 F/10V capacitor is mounted to ensure a ripple-less DC output voltage for the heaters. The SB140 Schottky diode has a 0.25V drop across it so the filament receives about 6.3 Volts. Easier to swap tubes now as you can use both 300mA and 365mA tubes without any problems. The crosstalk and distortion above a certain output l evels is now gone and the amp can make bigger output voltages as a bonus (60% higher output power !) More details about this modification are on the next page. The schematic of the heater power supply modification that must be added is shown on the left.
MORE details on the heater power supply modification: On the left is shown how the TRACO POWER TSR 1-2465 switch mode power regulator is to be connected to the other parts. ALL of the wiring (with the exception of the diode) has to be as SHORT as possible. This is to prevent switching noise from the regulator to interfere with the audio signals. If this is not implemented properly a small H.F. components may become present in the output signal.
The 100H inductor must be connected to the '+' solder joint of the big power capacitor. The middle (ground) pin of the regulator must be connected directly to the '-' of the power capacitor. Between the input (the left connection near the dot) and the ground (the middle connection) a small capacitor must be mounted. The value may be anything between 10nF and 100nF. In this amp 100nF SMD capacitors were used, the more common ceramic capacitors are shown in the drawing. Parallel to this small capacitor an electrolytic capacitor of 10 F/50V is placed again with as short as possible wiring. Values between 10F and 47 F can be used as long as the voltage rating is 35V or higher. Between the output pin (on the right) and the ground pin a similar small capacitor is placed again as close as possible to the regulator. An electrolytic capacitor of 470 F/10V is connected in parallel to this small capacitor. MIND the + and – connections of both electrolytic capacitors ! The value of this capacitor may range from 100 F to 470F and the voltage rating can be as low as 10V. The anode of the SB140 (or equivalent 1A Schottky diode with a similar low voltage drop) is connected to the output of the regulator which provides 6.5 Volts. A Schottky diode has a lower voltage drop (0.25V) at this current then a normal silicon diode such as 1N4001 has (0.55V at this current). The cathode of this diode is connected to the heater pin of the tube (pin 4). The 6V-12V switch still works as it is supposed to. All 'floating' parts have been secured with 'hot glue'.
Adjustment of the Anode voltage: Adjust both anodes of the tube to 13.5V, this is the optimal Anode voltage to obtain maximum output level and symmetric clipping after this modification has been done and IRL510/530s are fitted. Adjust the voltages about 5 minutes after start-up and check/re-adjust again some 10 minutes later. The adjustments have to be made with the small pots on the PCB. These pots can be reached from the bottom (small screwdriver) and top side of the PCB (bigger screwdriver) Pin 9 = ground ('-' lead of voltmeter for all measurements) Pin 1 = Anode 1 (left adjustment pot, '+' lead of voltmeter) Pin 6 = Anode 2 (right adjustment pot, '+' lead of voltmeter) Pin 4 = heater voltage (should be between 6.2 and 6.3V, '+' lead of voltmeter)
Important notes: In this amp the blue LED under the tube is connected with one leg to the ground plane as can be seen on the picture on the previous page bottom right. If this is done it is easier to get proper voltage readings when trouble shooting with the tube out of it's socket. The blue LED can also remain connected as it was but has some consequences when testing without the tube installed. When the tube is out of it's socket the red and blue LED will be 'off' when the amp is powered. Also when the tube is out of it's socket and one tries to measure the heater voltage values of around 20V will be present instead of the expected 6.3V. This is caused by current coming from the +24V flowing through both LED's and the 1k5 resistor. The SB140 blocks this higher voltage so it can't reach the TSR-2465 so nothing gets damaged but the glowing voltage will appear to be MUCH higher then the expected 6.3V. Only when the tube is inserted the proper glowing voltage can be measured. Also both LED's will be functioning again. In case the TSR gets very hot and the heater voltage is much too low check if the wire underneath the 2 big 68 Ω resistors has been removed. If it's not it shorts the heater and cause the TSR to go in overload. On the picture on the right the absence of the 3W/68 Ω resistors and the wire underneath the bottom resistor can be seen.
Alternative heater supply: Instead of the TSR2465 other (switching) regulators can be used or a linear power supply made with an LM317 for instance. Sufficient cooling (big heat sinks) must be applied in this case. The modifications above must all be carried out but the TSR circuit can be left out. Adjust this power supply to 6.3V. Connect it to pin 4 and pin 9 of the tube.
Class-A current adjustment: 2 resistors can be seen below the 2k bias pots. In this amp the 22 Ω and 10 Ω are in parallel and combined they are 6.8 Ω. Originally there are only 2 resistors in those 4 positions (so 2 positions are open). They determine the class-A current of the MOSFET end stage. The original value is 4.3 Ω and sets each channel to 280mA idle current. The reason the G2 is running so hot is because of these high currents. For single-ended class-A designs as this one it is mandatory for the bias current to be higher then the highest possible load current that may occur. This current is determined by the (maximum) output voltage of the amp and the minimal load resistance. This resistance is determined by the output resistor (30.1 Ω) + the lowest allowed headphone impedance. In the specs it states 12 Ω as minimum so 42 Ω total resistance. Since the output voltage of the amp is even higher with mod 7 as without it, the peak output voltage is 7.5V so with 42 Ω that means 180mA. When a 10% safety margin is imposed, about 200mA of class-A current is more then sufficient and with more realistic headphone impedances of 16 or 32 Ω this is WAY more then enough. The 4.3 Ω resistors can thus be made bigger in value and thus decrease the class-A current and as a result the heat coming from the LM317's and MOSFET's will be less. A 6.8 Ω resistor will give a 180 mA class-A current and for 200mA you will need 6.2 Ω. Total dissipated power of the whole amp will go from 13.5W to 9.5W. Without mod 7 the output voltage is 4.3V RMS max. With mod 7 the output voltage becomes 5.3V RMS max. Thus increasing the dynamic range and maximum output power of this amp by 2dB. Higher Ohmic headphones (120 – 600 Ω) will welcome the extra undistorted output voltage.
When all these mods have been done the bonuses are: 10x greater bandwidth translating to: much improved attack and no rolled off highs. 60% higher output power. Much less distortion at higher output levels. NO more crosstalk. warms up faster after switch on. Better tube roll selection capability (300mA and 365mA tubes correctly powered). Less heat coming from the amp. Better longevity. Schematics of all the mods described above:
The power switch was moved to the front and another volume knob was fitted. Also it got some really nice feet thanks to Mick. The G2 was kindly donated by Mike. Solderdude (Frans) a Rock Grotto member
