The Adcom GFA-555MKII is a completely upgraded version of the famous GFA-555 projected by Nelson Pass. Its retains the power output of the original GFA-555 and reduces distortion when driving difficult loudspeaker loads.

 Features:

• 200W continuous power into 8 ohms and 325W into 4 ohms
• low distortion THD < 0.04% from 20Hz to 20kHz
• high damping factor >800
• can drive loudspeakers with very low impedances and highly reactive
• very high current peaks into low impedance (exceeding 60 ampere per channel)
• TO-3 type bipolar output transistor with triple Darlington driver output stage configuration
• power supply with 1500VA toroidal transformer and 60000uF of filter capacitance

Though the Adcom GFA-555MKII is globally well projected amplifier, it has some flaws.

Circuit analysis

This amplifier employs a discrete differential Class-A front-end followed by a Class-A voltage gain stage (VAS) which amplify the input signal to the voltage required at the output of the amplifier. This high voltage signal drives the high current triple Darlighton follower output stage which amplifies the current by a factor of about 50000.

The input signal are AC coupled by C101 and filtered through network C102 and R103. The small signal transistors Q101 and Q102 form the differential input stage. Open-loop gain is defined by R105 ancd the bias current through Q101 and Q102. The small -signal gain is approximately 825/(2x25)=16. The next voltage gain stage consists of Q107 with Q108 as a current-source load. DC biad is set by R116, D103 and D104. Open-loop gain is defined by R112 and R113, with R201, R301, C105, C201, and C301 providing high frequency compensation.

Feedback is provided from the output to the base of Q102 by the network R123, R124 and C106. C106 provides a high frequency roll-off above 200kHz, improving stability by taking high frequency feedback before the triple Darlington.

The amplifier has a DC coupled global negative feedback and this requires a DC error correction provided by IC101. Any DC imbalance in the amplifier is corrected by R125, R126, R127, C107, C110 and IC101. Any DC error at the amplifier output is servoed back through IC101 to adjust the DC current through the input transistor.

The bias network of R117 through R119 and Q307 form a temperature-compensated DC bias voltage to the input of the triple Darlington follower output stage. Mid and high frequency bypassing is provided by C104.

R901 and C901 provide a load for the amplifier at high frequencies, stabilizing the amplifier under varying load conditions.

The output stage consists of two sets of 4 parallel transistors operated as emitter followers, driven by another pair of emitter followers. This configuration minimizes distortion caused by varying load impedances. The output transistors have 0,22 ohm ballast resistors to ensure current sharing and bias stability.

General flaws:

• absence of speakers protection
• wrong sizing of VAS transistor
• absence of heatsink for VAS transistor
• absence of power supply bypassing capacitors for input stages circuitry

UPGRADES

In this article I illustrate how I have been transformed this amplifier into a very High End amplifier with high improvements in sound quality and general stability.

Main power supply improvements

The Adcom GFA-555MKII has a generous power supply with a 2x 25 A bridge rectifier and 4x 15000uF capacitor filtering. I have changed the original 25 A KBP2504 bridge rectifier with 50 A FB5006 bridge rectifier. I have changed the 100nF bypassing capacitors with new WIMA MKS2 type and I have changed the capacitors discharge resistor with a 4.7 kΩ.  I have added a 10nF 380V X2 type capacitor on the AC side of bridge rectifier for filter high frequency noise on AC.

List of changes
Description	Quantity	Original Parts	New parts
D801, D802	2	KBP2504 25A 400V	FB5006 50A 600V
C806, C807, C808, C809	4	100nF 100V	100nF 250V WIMA MKS2
R801, R802, R803, R804	4	3.9 kΩ 3W Metal Oxide	4.7 kΩ 5W Metal Oxide
Filter on AC side of bridge rectifier	2	Not present	10nF 380V X2

 

Power supply improvements on input stage board

The Adcom GFA-555MKII don't have a dedicated bypassing capacitors on the input stages board. I added a couple of capacitors for rail, one 120uF 200V Nichicon Aluminum Electrolytic type and one 47nF MKM4 250V WIMA type (you can use also 100nF). This is the first and simple upgrade that create a bypass filter between the main power supply and the rails that comes to the input stages board.

This improve power supply filtering performance and reduce ripple and noise.

List of changes
Description	Quantity	Original Parts	New parts
For bypassing power supply of input stages board	4	Not present	120uf 200v Nichicon Aluminum Electrolytic Cap
For bypassing power supply of input stages board	4	Not present	47nF MKM4 250V WIM

Power supply improvements on output stage boards

The output stages have bypassing capacitors C202, C302, C252, C351. These capacitors, with normal use and after several years, tend to bulge. In my sample of amplifier I have founded them bunged and fault. 

I have replaced them with 120uF 200V Nichicon Aluminum Electrolytic type and I have added a 22nF 600V Sprague Orange Drop Polypropylene Film type (you can use also 100nF).

This improve power supply filtering performance and reduce ripple and noise.

List of changes
Description	Quantity	Original Parts	New parts
C202, C302, C252, C352	4	47uF 160V Electrolytic Cap	120uf 200v Nichicon Aluminum Electrolytic Cap
parallel with C202, C302, C252, C352	4	Not present	22nF 600V  Sprague Orange Drop Polypropylene Film

 

New components on input stage board

I have changed all transistors and all capacitors on the input stage board except for C103 and C153. I have changed the trimmers R119, R169  (for bias setting) with a multiturn precision type. I have matched all transistors with the same Hfe.

List of changes
Description	Quantity	Original Parts	New parts
Q101, Q102, Q151, Q152 (input stage)	4	2SC2362
Q103, Q153, Q307, Q257	4	2SC2240
Q104, Q106, Q154, Q156( PNP)	4	2SA1016
Q105, Q107, Q155, Q157 (VAS PNP)	4	2SA1210 / 2SA1142	2SA1837
Q108, Q158 (VAS NPN)	2	2SC2912 / 2SC2682	2SC4793
D101, D102, D103, D104, D151, D152, D153, D154	8	1SS178	1N4148
C101, C151 (input signal)	2	1uF 100V MKC	1uF WIMA MKP4 250V
R119, R169	2	Trimmer 2 KΩ	Multiturn Trimmer 2 KΩ
C104, C154	2	4,7uF 50V	4,7uF 63V ELNA SILMIC II
C107, C108, C110, C157, C158, C160	6	100nF 50V	WIMA MKS2 0.1uF 250V
C102, C152	2	330pF 100V mica	300pF 500V silvered mica
C105, C155	2	15pF 500V mica	22pF 500V silvered mica
C106, C156	2	33pF 500V mica	33pF 500V silvered mica
C109, C159	2	470uF 25V	470uF 35V

New VAS transistors and addition of heatsink

The transistors of the voltage amplification stage (VAS) are Q107, Q157, Q108, Q158 and are not well sized. The original parts are 2SA1210/2SC2912 and they have maximum collector-emitter voltage of 200 V. If the voltage selection on primary of transformer are set for example to 220 V and I have a voltage of 230 V on line of electrical  distribution such as in some european country as Italy, the DC voltage that I have measured on power supply rails is 86.8 V. This implies that the total amount of DC voltage between collector and emitter of VAS transistors are 2x86.8=173.6 V, very near to 200 V considering the Safe Operating Area  of the 2SC2912. In my sample of amplifier I have found that the transistors mounted are 2SC2682 that have a VCEO=180 V that is very low value considering a working voltage of 173.6 V.  This causes the reduced life of transistors and a probable failure.  I don't know why Adcom has installed this type of transistors but for me it's not  a good choice.

The new transistors that I have mounted are Toshiba 2SC4793 (NPN) and 2SA1837 (PNP) (note that these transistor have inverted pin compared to the original 2SA1210/2SC2912). These transistors have higher collector-emitter voltage specification , higher current and higher power dissipation.

Transistors of the Voltage Amplification Stage (VAS)
Characteristics	Symbol	Original Parts 2SC2912 (NPN) 2SA1210 (PNP) Rating	New parts 2SC4793 (NPN) 2SA1837 (PNP) Rating
Collector-base voltage	VCBO	200 V	230 V
Collector-emitter voltage	VCEO	200 V	230 V
Emitter-base voltage	VEBO	5 V	5 V
Collector current	IC	0.14 A	1 A
Collector power dissipation at Ta=25°C	PC	1 W	2 W
Collector power dissipation at Tc=25°C	PC	10W	20 W

List of changes
Description	Quantity	Original Parts	New parts
Q105, Q107, Q155, Q157 (VAS PNP and current generator)	4	2SA1210 / 2SA1142	2SA1837
Q108, Q158 (VAS NPN)	2	2SC2912 / 2SC2682	2SC4793

Another well know problem of this amplifier and a cause of very large fault is the overheating of VAS transistors. I don't know why Adcom don't have projected to mount the heatsinks on VAS transistors but for me it's a very serious flaw of this amplifier. Without heatsink the transistors of VAS stage are very hot, (you can not touch them) and this compromise the entire stability of the amplifier. The overheating of these transistors are demonstrated  also by the browning of solder side of PCB.

I have projected and made a custom heatsink for VAS and relative current source transistors. With this mod the temperature of these transistor remain around 51 °C, a very good result. These transistors now can work in their Safe Operating Area, do not regenerate a temperature dependent noise and have more margin of stability. The addition of the heatsinks is a very important upgrade for this amplifier.

New components on output stage boards

I have changed all transistors with new more performant types. I have replaced all old Toshiba transistors with new On-Semi matched pair transistors. The new transistors have much more Safe Operating Area and more current capability, this improves the capacity of drive difficult loads and increases the stability at high temperature.

An important improvement is to change the original resistors R901, R951 with a new 5W Metal Oxide type. In my sample the original resistors was broken and the amplifier went into oscillation.

List of changes
Description	Quantity	Original Parts	New parts
Q201, Q251 (predriver NPN)	2	2SC2912 / 2SC2682	MJE15034
Q202, Q252 (driver NPN)	2	2SD1047	NJW3281G
Q203, Q204, Q205, Q206, Q253, Q254, Q255, Q256, (final output stage NPN)	8	2SD424	MJ21194
Q301, Q351 (predrivers PNP)	2	2SA1210 / 2SA1142	MJE15035
Q302, Q352 (driver PNP)	2	2SB817	NJW1302G
Q303, Q304, Q305, Q306, Q353, Q354, Q355, Q356 (final output stage PNP)	8	2SB554	MJ21193
R901, R951	2	10 Ω 2W	10 Ω Metal Oxide Film Resistors 5W
C901, C951	2	100nF 100V	100nF 630V MKP SCR
C201, C251	2	68pF 500V mica	68pF 500V silvered mica

 

Addition of speakers protection

The Adcom GFA-555MKII doesn't have a speakers protection circuit and doesn't have an Anti-Bump on power on.

I have added a speakers protection circuit that have also an Anti-Bump function. This circuit is based on UPC1237 integrated chip and you can find it also on Ebay. For power supply  this circuit I have used a secondary rail of the main transformer that is originally dedicaded for optional fan motor. This rail has 22V AC and I have build a little circuit with a bridge rectifier and a capacitor to obtain a voltage of 31V DC for supply the speakers protection circuit.

Other mods

I have installed an IEC power connector with a little toroidal ferrite filter and I have changed the speaker connectors.

Photo after the Upgrades

     

Our Services

We offer repair/restore/upgrade services, please contact us to get a quote.

• Prev

Category: Upgrades